Amazon.com Aktie

Thank you for standing by. Good day, everyone, and welcome to the Amazon.com Q3 2026 Financial Results Teleconference. [Operator Instructions] Today's call is being recorded.

For opening remarks, I'll be turning the call over to the Vice President of Investor Relations, Mr. Dave Fildes. Thank you, sir. Please go ahead.

Hello, and welcome to our Q3 2026 financial results conference call. Joining us today to answer your questions is Andy Jassy, our CEO; and Brian Olsavsky, our CFO. As you listen to today's conference call, we encourage you to have our press release in front of you, which includes our financial results as well as metrics and commentary on the quarter. Please note, unless otherwise stated, all comparisons in this call will be against our results for the comparable period of 2025.

Our comments and responses to your questions reflect management's views as of today, October 29, 2026, only and will include forward-looking statements. Actual results may differ materially. Additional information about factors that could potentially impact our financial results is included in today's press release and our filings with the SEC, including our most recent annual report on Form 10-K and subsequent filings.

During this call, we may discuss certain non-GAAP financial measures. In our press release, slides accompanying this webcast and our filings with the SEC, each of which is posted on our IR website, you will find additional disclosures regarding these non-GAAP measures, including reconciliations of these measures with comparable GAAP measures. Our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our results are inherently unpredictable and may be materially affected by many factors, including fluctuations in foreign exchange rates and energy prices, changes in global economic and geopolitical conditions, tariff and trade policies, resource and supply volatility, including for memory chips and customer demand and spending, including the impact of recessionary fears, inflation, interest rates, regional labor market constraints, world events, the rate of growth of the internet, online commerce cloud services and new and emerging technologies and the various factors detailed in our filings with the SEC.

Our guidance assumes, among other things, that we don't conclude any additional business acquisitions, restructurings or legal settlements. It's not possible to accurately predict demand for our goods and services, and therefore, our actual results could differ materially from our guidance.

And now I'll turn the call over to Andy.

Thanks, Dave. We're reporting $201.9 billion in revenue, up 12% year-over-year. Operating income was $28.1 billion, up 61% year-over-year. Q3 was another strong quarter for Amazon. In AWS, our plan is working. Despite lapping a comparable year-over-year period where we added all of our capacity, AWS grew 39.3% year-over-year, accelerating for the sixth straight quarter, our fastest growth in nearly 5 years. Backlog is $578 billion, growing triple digits year-over-year and well diversified.

We're installing an AI accelerator about every 14 seconds and almost as fast as we install capacity, customers are gobbling it up. Our core compute and storage revenue grew at their fastest rates in several years. Our operating margins have long been strong on our core offering, but rapidly improving on the AI side, too, well ahead of where they were on core at a similar time of evolution.

Even with heightened component costs like memory, the significant depreciation from the CapEx we're investing. We continue to accelerate our CapEx payback and have even more conviction than we already had that the investment we're making in AWS will yield a much larger revenue and free cash flow generating business than we would have imagined a couple of years ago with strong return on invested capital.

We have a compelling flywheel spinning in AWS. Borrowing from Jim Collins [indiscernible] definition and good to great, where if you do a, you almost can't help with if you do B, if you do B, you almost can't help with do C and so on.

In AWS, having the leading functionality securing and operational performance, can't help but drive customers to run their apps and store their data in AWS, which can't help but drive AWS AI demand because customers prefer their production inference to be near the rest of their data and apps, which can't help but drive core AWS demand because AI uses core for work like reinforcement learning, post-trading tool usage, which can't help, but drive reinvestment back into our broader AWS functionality.

This is a different model in primarily hosting GPUs or renting short-term data center capacity. This is a long-term platform strategy that broadly helps customers advent and evolve their customer experiences, operate stably and build a reliable partnership they can grow for many years. And it also yields very attractive long-term economics for us regardless of supply scarcity.

There are many examples that illustrate our progress in AWS, but I'll share a few. On the core side, it's important to remember that the vast majority of global IT spend, about 85% is still on premises, and we're in a 10- to 20-year cycle where that equation will flip to the cloud. Conventional wisdom has been that lynchpin workloads like mainframes or storage array networks would never move to the cloud. But these workloads are moving as AWS has built unique cloud substitutes and AI-driven migration capabilities to make this transition much easier. And enterprises get the benefits of less CapEx, lower operating costs, faster innovation, better developer productivity and data accessibility to more easily leverage AI.

Our core business has attractive margins because we've garnered very efficiently over the last 20 years. As our core business grows, we keep reinvesting in our core AI functionality. For example, as customers generate increasingly more data and seek to get value from an analytics and AI the complexity of moving data between data stores and analytics engines is onerous and expense. With our acquisition of DuckLabs, DuckDB will become the universal Sequel dialect of AWS storage analytics, simplifying data processing with a modern analytics engine that is fast, simple to use and ubiquitous and where customers will be able to write SQL queries once and then execute using any AWS analytics service.

For the first time in analytics, customers will have 0 migration or portability costs between query engines. This is a big deal. And you can imagine how it will also help agents do efficient querying as well.

On the AI side, we continue to expand the capabilities of our inference service, Amazon Bedrock, and it's leading to remarkable growth with Fortune 500 companies spending nearly 7x more on Bedrock year-over-year and total Bedrock token volume growing nearly 17x year-over-year. Bedrock is the fastest growing multibillion-dollar service in AWS' history. We recently added open AI models to Bedrock and usage is growing quickly. We also just added Bedrock managed versions of Kimi, GLM and SpaceX models.

You can't have a conversation about AI today without talking about security. It's another reason customers keep choosing Bedrock. We've designed Bedrock for customers' data remains secure and private. Inputs and outputs are never shared with model providers or used to train based models. When customers tune a foundation model, we create a private copy specifically for their use, and data is fully encrypted both in transit and arrest.

As frontier models get more powerful, they're servicing software vulnerabilities previously not found and at unprecedented rates. We built AWS continuum to discover, prioritize, validate and remediate code vulnerabilities and machine speed. It's already being used by customers like Capital One, IBM, Infor, Luminate and Robinhood.

In Q3, we announced that we'll integrate it into the tools where codes being written in Anthropic's Claude Code, OpenAI's Codex, and AWS's Kiro. The interest from customers exceed anything we've ever seen for a new security service.

Agents continue to be deployed at accelerating rates, many of which are being built on Bedrock agent core building blocks. Tens of thousands of customers currently use AgentCorp, deploying 2x more agents within Q3 and than the prior 3 quarters combined. In addition to building their own purpose-built agents, most companies will also use turnkey agentic services, and we offer several of the customers love. Kiro for coding, Amazon Connect for contact center agents and Amazon Transform for migration. They're all iterating quickly in picking up customer momentum.

As our intelligent AI work companion, which is growing very rapidly as large enterprises like Mondelez, EchoStar, 3M, Southwest Airlines and DXC are choosing Quick because it's purpose-built for work with no compromises on customer data.

Finally, our chips business continues to grow triple-digit percentages year-over-year and revenue commitments for Graviton and Trainium have each increased more than 20x year-over-year. AI continues to be a tailwind for Graviton because post trading reinforcement learning and agent tool use is mostly done on CPUs versus AI accelerators and Graviton is the strongest CPU chip, offering up to 30% to 40% better price performance than alternatives.

Customers using our custom silicon saved more than $2 billion in Q3 alone, and we just opened source torch neurons as developers can run existing pie torch code, which most already use on Trainium without rewriting a thing. As we launch new generations of chips, demand for prior generations remained strong. After our launch of Graviton 5, backlog for prior generation Graviton instances still more than doubled quarter-over-quarter. And short-term capacity for prior generation training instances sell at a premium compared to the first half of the year.

With training demand so strong, customers continue encouraging us to make it available outside of AWS we intend to do so late in 2027. Overall demand for AWS capacity remains exceptionally strong. Enterprises are still very early in using inference of scale, and yet inference already accounts for well over half of AWS' AI revenue and growing. And we have a lot of training revenue as well. We still expect 2026 cash CapEx of approximately $220 billion, the majority for AWS and expect to spend meaningfully more in 2027, with a clear line of sight to attractive financial returns, as I laid out on our last call.

AWS is a $184 billion annualized revenue run rate business, and yet the opportunity remains extraordinary. With our continued pace of innovation, new core and AI functionality, enterprises wanting to run their core and AI workloads in AWS and the amount of capacity that we're adding each year. We believe that AWS can grow north of 40% in each of 2027 and in 2028, and be a $1 trillion revenue business in the next 5 to 10 years. We're in the very early days of this inflection our strategy is working, and I expect it will for the foreseeable future.

In stores, I think some folks are sleeping a bit on how much progress that business has made and what has become in the last few years. I'm not aware of another retail business approaching Amazon's revenue size, hundreds of billions of dollars, with double-digit operating margins inclusive of [indiscernible] is growing at this pace. Year-to-date, units have grown nearly 15% year-over-year, and that's on an enormous base having sold more than 30 billion items last year. In absolute dollar terms, we've added more than $90 billion in gross sales in the trailing 12 months. And we're growing meaningfully faster than other retail competitors, adding about double the gross sales per quarter than our next largest competitor in the U.S. and consistently taking retail market segment share.

Over the past few years, our North America and international segments have each expanded operating margins by nearly 10 percentage points, which enables us to reinvest in making the customer experience even better. Remember, it's still early. The significant majority of retail, nearly 80% currently occurs in physical stores. That will eventually flip to e-commerce over time. We're helping enable this shift by rapidly building out our efficient and speedy fulfillment network, providing delightful customer experiences while driving significant free cash flow and return on invested capital.

Each of our priorities from faster delivery, to cost to serve, to groceries and pharmacy, to AI shopping are force multipliers, and we're firing on all cylinders. Our delivery speeds keep getting faster as we invest in our same-day network as well as our Amazon Now and Amazon Prime Air ultrafast services. We had our fastest Q3 ever globally, and we delivered more than 1.5 billion items to prime customers same-day or overnight in Q3. We're energized by the customer response to our rapid rollout of AmazonNow around the world, is now available in over 450 towns and cities, and the number of items ordered through AmazonNow is up 10x year-over-year.

We're on track to reduce cost to serve per unit for the fourth consecutive year, which enables us to continue meeting or beating competitors' prices and cost effectively deliver a broad selection of fresh groceries and everyday essentials. Increasing speed is fueling our grocery business, including perishables and everyday essentials and gross sales grew to nearly $250 billion in the trailing 12 months. In Q3, fresh groceries and everyday essentials grew about 3x faster than the rest of our store, representing about half of all units sold. Customers who shop perishables with same-day delivery by nearly 7x more everyday essential units than those who don't. An active grocery customers shop nearly 4x more non-grocery categories like home and apparel and spend over 2x more on those categories compared to nongrocery customers.

Amazon Pharmacy is also growing fast. The number of customers using it is up over 60% year-over-year with growth accelerating each quarter through 2026, and repeat customers grew more than 50% year-over-year. J.D. Power recently ranked Amazon Pharmacy, the top mail-order pharmacy for the third year in a row, and we keep making it even better, now offering 3-hour prescription delivery more than 350 cities.

Over 400 million customers have used Alexa for shopping, our agentic shopping system in the last 12 months with customer interactions up 50% year-over-year. U.S. customers using Alexa for shopping are spending 50% more per order than those who don't. It researches products compares options, checks a full year of price history, get personalized gift recommendations, and tracks deals and price drops on customers' behalf. Though we anticipate working with horizontal agents, we believe Alexa for shopping is going to be the primary shopping agent customers use to shop Amazon.

[indiscernible] shopping assistant constantly by your side as you browse, get inspired, have questions on product detail pages and that you can simply ask to buy something based on your prior purchases that already has your history, your shipping address and payment info is uniquely helpful and convenient for customers.

Moving to Amazon Ads. We generated $19.6 billion of revenue in Q3, up 12% year-over-year. Though there were some timing anomalies in this past quarter, we continue to be pleased with the overall momentum across our ad offerings. I'll mention 3 items impacting the growth rates. First, we shifted Prime Day into Q2 for most countries, moving significant sales and ad revenue that were in Q3 '25 into Q2 '26, a tailwind to our Q2 2026 rate and a headwind to Q3 2026. For advertising, that was 850 basis points sequential impact to Q3 year-over-year growth.

Second, Q2 '26 included the first season of NBA on prime, which last year's Q2 didn't have, while Q3 does not include any NBA it starts in Q4.

And third, in Q3 2025, we changed our search engine marketing programs to drive more traffic to search pages on Amazon. It's a better experience for customers, and we also monetize that traffic at a higher rate and helped accelerate advertising revenue growth starting a year ago, and we started lapping that in Q3. We expect year-over-year growth will return to the high teens in Q4.

A natural question is whether agentic shopping from external horizontal agents is meaningfully impacting Amazon Ads. And the answer is no. Few customers are doing much shopping this way yet as the experience is still not great nor more convenient than coming to a retailer directly to research or 1 click order there. Q3 was chockful of innovation from our ads team that we expect to have impact in the coming months, have valued by Amazon Ads Agent, which brings together our full advertising offering under a single AI-first platform, backed by trillions of shopping, browsing and streaming signals, it supports advertisers at every step with insights, answers and the ability to optimize and act on their behalf.

We also launched full funnel campaigns in Amazon Ads Agent, bringing sponsored ads, display, video and streaming TV into a single campaign. Advertisers provide initial inputs, including budget and duration and Amazon's AI plans, executes and continually optimizes the rest.

We're continuing to see momentum in several other areas. I'll mention just a few. Prime Video has emerged as a significant business with a revenue run rate in the tens of billions of dollars now profitable in its own right. We have momentum with hit shows like Carrie, The Summer I turn Pretty, The Boys, Off-Campus, Reacher and Fallout. And breakout movies like Project Hail Mary, along with compelling live sports, including Thursday night football off to its best start ever in 2026, NBA on Prime, NASCAR, The Masters and UEFA Champions League soccer.

We recently launched our most powerful devices lineup yet with 34 new products across Echo, Kendall, Alexa tablets, Fire TV, Ring and B. It's early, but we're off to a very strong start and excited for more customers to get their hands on these as we head into the holidays.

And finally, Amazon Leo remains on track to begin initial service this year. We're heading into 1 of our most exciting stretches of the year with both reinvent and the holidays, and I want to thank our teammates for their hard work and invention to make customers' lives better and easier every day. What we're building at Amazon is unlike anything that's ever been seen before. There has never been a business with $1 trillion of annual revenue. Amazon is on a path to have 2 of them in AWS and stores. It's rare, and we're energized by what's ahead.

Over to you, Brian.

Thank you, Andy. Let's start with our top line financial results. Worldwide revenue was $201.9 billion, a 12% increase year-over-year and 13% increase year-over-year, excluding the impact of foreign exchange. Worldwide operating income was $28.1 billion, with an operating margin of 13.9%, which is our highest operating margin ever.

Starting with our North American International segment results. First, let me remind you that Prime base shifted this year with the sales event occurring in Q2 for most of our largest countries, including the United States. In 2025, Prime Day was entirely in Q3. Excluding the impact of this Prime Day time change, third quarter 2026 year-over-year growth would be approximately 400 basis points higher. In Q3 on customer engagement and a continuation of trends from the first half of the year. In the North America segment, third quarter revenue was $112.9 billion, an increase of 6% year-over-year. International segment revenue was $43.1 billion, an increase of 8% year-over-year, excluding the impact of foreign exchange. Worldwide paid units grew 12% year-over-year.

Prime remains a key pillar of our business. Hundreds of millions of Prime members are responding to our combination of vast selection, low prices and record delivery speeds. Over the past year, monthly orders of fresh groceries placed by our Prime members for 3 same-day delivery have more than tripled.

Let's shift to profitability. North America segment operating income was $8.7 billion, with an operating margin of 7.7%. International segment operating income was $1.6 billion with an operating margin of 3.7%. In our fulfillment network, the team navigated higher transportation costs from fuel inflation and higher linehaul rates. Even with these headwinds, we still lowered our overall cost to serve in Q3 through better inventory placement, shorter shipping distances and package consolidation. Shipping costs increased 11% year-over-year versus a 12% increase in units. Excluding fuel and line haul inflation, shipping costs grew 8% or approximately 400 basis points slower than unit growth.

Robotics and AI are delivering measurable gains through productivity and employee safety, which we'll expect to compound in the future. In 2026, we doubled our fleet of robotic arms like Cardinal and Sparrow. We're also developing robotics foundation models that enable faster programming and teach robots increasingly complex physical tasks with the goal of delivering useful robotics at scale. Looking ahead, we see meaningful opportunities to improve productivity across our global fulfillment network and lower our cost to serve while continuing to get products to customers faster. Our operating margin fluctuate quarter-to-quarter, but we're focused on sustained long-term improvement in our cost to serve per unit.

Moving to our AWS segment. Revenue was $4 million and grew 39.3% year-over-year driven by both core and AI services. We added $3.8 billion in quarter-over-quarter revenue. AWS now has an annualized revenue run rate of $184 billion. Customers continue to move to the cloud and scale their use of AWS for services. And increasingly, customers who want the full benefits of AI are accelerating their transition to the cloud.

We see a strong linkage between AI spend and growth in our revenue for core services. As customers invest in AI we see a corresponding increase in core consumption, and we expect that relationship to strengthen as more AI workloads move into production and drive additional demand for compute, storage and networking. AWS operating income was $17.8 billion with an operating margin of 38.7%, reflecting strong growth and our focus on driving efficiencies. Our investments in custom chips and custom networking along with software and process improvements are helping us optimize capacity and build a more cost-efficient infrastructure. As we've long said, we expect AWS operating margins to fluctuate driven in part by the level of investment we're making at any point in time.

Our AI offerings are gaining traction with customers with AI revenue growing triple digits year-over-year. Bedrock has been a significant growth driver.

Now turning to our cash CapEx, which was $75.9 billion in Q3. This primarily relates to AWS and AI as we invest to support strong customer demand. We continue to make significant investments, especially in as we believe it to be a massive opportunity with the potential to drive long-term revenue and free cash flow as well as return on invested capital.

I'll finish with our financial guidance for Q4. Q4 net sales were expected to be between $240 billion and $247 billion. We estimate the year-over-year impact of changes in foreign exchange rates based on current rates, which we expect to be a headwind of approximately 60 basis points in the quarter. Q4 operating income is expected to be between $28.5 billion and $33.5 billion. I'll mention 1 factor impacting our Q4 guidance. Amazon Leo's commercial service is on track to begin initial service this year, and we expect to begin capitalizing certain costs in Q1 of next year, including production and launch costs. This is later than I've been predicting in earlier earnings calls. As a result, we will continue to expense the majority of Leo related costs in Q4. We're ready to serve customers this holiday season and to support the millions of customers building with AWS.

I want to thank our teams across Amazon for their hard work and focus on customers. Our commitment to elevating the customer experience remains the only reliable way to create lasting value for our shareholders.

With that, let's move on to your questions.

[Operator Instructions] And our first question comes from the line of Justin Post with Bank of America.

Andy, continue to see the cloud business accelerate, which is really nice. But the quarter-over-quarter growth slowed, and I'm just wondering if there's any callouts there on the customer side or token costs or anything that kind of drove that? And how you're thinking about the capacity adds in Q4?

And then, Brian, maybe just any impact of fuel costs on your guidance in Q4?

Yes. I'll start with the growth question. We're really pleased with the growth, 39.3% on a business, it's got $184 billion annual revenue run rate is quite unusual. Our fastest growth in about 5 years or 6 quarter consecutively of acceleration. And we're adding quite a bit of capacity. I think if you look at what's driving the growth right now, it's a combination of a few things. First, on the AI side, the unusual top-to-bottom stack functionality we have, whether you're building your own model with SageMaker AI, which is the easiest way to build your own model, where do you want to leverage 1 of the leading frontier models with high security and very strong price performance, people can use Bedrock.

Whether you are building your own agents where we have capabilities like strands and agent core to build and to run those securely and scalably, or are you going to use turnkey agents, which most companies do, if you just look across Kiro for coding and Connect for call center and transform for migrations and Quick, which is growing incredibly fast as an AI companion work. People are adopting those very quickly.

And then we add a lot of capacity every quarter. Now the amount that we add every quarter is not equal. And we're adding different amounts each quarter. But if I look at how much we've added this year and what we have planned over the next, call it, 8 quarters or so, which is about as far out as we get 8 to 10 quarters, we have a lot of capacity coming as well. And then I layer on top of that, too, just the impact of chips. We're very well positioned for what's happening in AI because increasingly, what you're finding right now is most customers, every customer that runs that uses AI at scale. But particularly enterprises now as they're starting to deploy more inference into production, they are very eager to save on price performance and training them with a meaningful advantage it has in price performance versus other alternatives is very compelling to people.

And then as more and more people are using AI in a very broad way, it drives a lot of the core in capabilities like reinforcement learning and post-trading and tool calling and we have the leading CPU chip in Graviton which is about 30% to 40% more price performance than the other x86 providers. So it's -- those things are really driving very significant growth. The amount of capacity we add each quarter will vary, but we're -- I mean, at this scale, to be projecting growing over 40% year-over-year in each of 2027 and 2028 is pretty significant and the fact that we believe that this can be a $1 trillion revenue business in 5 to 10 years, it's an extraordinary, unusual opportunity that we're very excited to be pursuing.

And Justin, on your question about fuel and linehaul inflation, I would say -- I'm not going to -- I don't have an exact number for you, but it is higher in Q4 than it was in Q3. If you remember, we instituted a fuel surcharge at the beginning of the year for FBA sellers that recovers some, but not all of the costs that we're seeing on fuel for those shipments and then the remaining partners on the retail inbound shipments and outbound shipments. So it's in there. It's a little bit higher than Q3.

Our next question comes from the line of Brian Nowak with Morgan Stanley.

I have 2. Andy, your excitement about AWS is coming through the line. It's pretty rare you would make comments about $1 trillion of revenue 5 to 10 years from now. It's more than you get. So can you just maybe walk us through how you get to that $1 trillion of dollars? Is it top down? Is it bottoms up? So what gives you confidence that you're going to be able to put up that type of revenue growth and to get to that type of number in the next 5 to 10 years?

Then the second one is for either of you on the robotics side. As we look into '27, what are the key milestones you're both focused on the robotics side for the retail business?

Yes. On the AWS side, if you just look at the growth rates, Brian, look at the growth rates on the run rate and then look at what we're projecting in '27 and '28. And as I've said a few times, we will always -- despite how much capacity we're providing -- we -- the big chunks of capacity that people take, the hundreds of megawatts or sometimes up to 1 gigawatt of capacity, those big chunks are reserved well in advance because you can't leave a gigawatt of capacity hanging around. And so we have very good visibility into '27 and '28 and some of '29 too with respect to what customers want in those really big chunks of capacity.

We will always leave a fair bit of capacity available for on-demand on EC2 or in Bedrock, and we're working closely with our customers and they'll have plenty of capacity to be able to use in AWS. But we have very good visibility into the next few years and what AWS looks like, those types of growth rates when you do the math on the revenue and the revenue run rate we have get you pretty quickly to escalating numbers.

And then I'd say that on top of not believing that a lot of the usage we see today is going to attenuate any time soon. We're still really early. Enterprises using inference in their production applications in a pervasive way. It just -- it is just really early. And some of that is enterprise is having enough people who know how to operate in AI and post trade and fine-tune models for the right applications. Some of it is the cost of inference.

I think if we all want enterprises to be able to use inference the way we all believe they will, we have to keep working on taking the cost of infants down. That is a huge mission for AWS. It's one of the significant reasons why we're building our own custom [indiscernible] that we just really believe it's incredibly important for customers to be able to use inference in a much more cost-effective way. That has come down a lot, the cost of inference over the last couple of years, and I think it will come down meaningfully more.

And so -- and then you just think about what you're seeing. If you think about security, as an example, just by itself, the models today are so advanced that the vulnerabilities that they're finding humans just haven't felt. And if you've had access to these models, it leads to -- you're using these models to both find them and ultimately to help with patching them and deploying. And there are some people who think that the amount -- the total amount of tokens used over the next year or 2, 20% of it, if some significant percent of it may end up being just security, and that's a relatively new use case on top of what we're seeing with coding and how that continues to grow on top of what we're seeing with the various, I'll call it, AI companions, both in the consumer space and in the workspace, like with Quick. And these are just the initial big use cases.

And you layer on top of that, what you're going to see with enterprises where I believe virtually every enterprise -- modern enterprise production application will be rewritten to be agentically capable because that's the way you're going to have to deploy if you want to move quickly or you want to actually garden the right way against some of the security risk there is just so much AI coming. And so that's -- you can't predict with any certainty 5 to 10 years from now, especially these types of growth rates, but it's not a stretch for us to imagine that this could be over a $1 trillion business.

On the robotics side, what I would say is that it's pretty remarkable what's happening in robotics right now. I'll say in general, but particularly at Amazon. We have over 1 million robots now in our fulfillment network. And while they're having very significant impact on our cost structure and on our productivity and on safety, if you look at our safety numbers and our fulfillment network, there are record low numbers. We've made huge progress over the last 4 to 5 years, and robotics has been a big piece of that.

And yet I would tell you, we're still early with respect to both what we have deployed -- built and deployed and what we're working on. And I think that you'll see, we will always have a lot of teammates in our fulfillment centers. But I think a lot of the jobs that we have that are more repetitive that we have humans doing will be done by robots. I think we will -- we're investing in all sorts of additional robotics of all different form factors. I think we have the opportunity to also assist on some of the delivering of the last 100 yards to customer stores.

We have a lot coming. And so -- for us, the way we measure robotics are, does it lead to comparable or better safety, cost structure and time to delivery. And when we like that combination of economics, we will deploy it at scale in our fulfillment network. And if I look at the new designs that we have coming in our fulfillment network, they -- including our expansion of same-day facilities, they are very heavy on robotics. So we think it had a big impact, and we think it's small compared to what it's going to be in the next 5 years.

Our next question comes from the line of Mark Mahaney with Evercore ISI.

Two things, one. First, thanks a lot for the forward comments on AWS growth. Could I ask you to draw out margins, too. I know that margins can move up and down quarter-to-quarter, but over the next 2 years, do you think the current margin structure is relatively stable. We kind of hold at these kind of high 30s, 40% operating margins for AWS. And then on perishable checkout, I'm super excited about this opportunity just watching how it's rolled out in my neighborhood and in other areas.

How much coverage do you have for perishable checkout and others, how many time households do you think you've covered with perishable checkout? And for those that have used it, how much of a boost do you think that is per prime household? Like is this something that increases spend per prime household by 20%, 50%, 100%? Just given how large the fresh grocery category is?

Mark, let me start with the operating margin outlook for AWS. So we're not guiding up beyond -- we're not guiding at all on our margin for AWS. But I can tell you -- give you my color on it. So very -- it seems very strong op margin, 38.7% this quarter. And I would add that the profitability you've been seeing recently is really not a random out, but it's a result of a lot of disciplined efficiency gains, capacity optimization, closely managing our fixed cost. And as Andy said, the inclusion of over time, more and more custom chips into our cost structure. However, there are other things bouncing around. In fact, the outsized growth of AI puts margin pressure on the business as a whole because the margins are lower there.

We are happy with the progression in both core and AI margins. And there -- as Andy said earlier, they're ahead of where the core was at a similar point in time, but they're still lower. So there's going to be that mix issue. There's also headwinds from component pricing that we've talked about. So cannot predict it, but there's going to be always a confluence of these multiple factors that we're working very hard on the cost side and the efficiency side and dealing with investments and new product margins as they start up and launch and reach scale.

On the perishables checkout, Mark, I share your enthusiasm about it. It's -- as you know, we've had a big grocery business for a while. And it's mostly been in the middle aisle what you'd find a grocery store, consumables, canned goods, pharmaceutical items, beauty products. And we've had this leading organic grocery store in Whole Foods Market, which also is growing faster than the rest of the grocery segment, both on the top line and the bottom line with a new format and daily shop, and urban environments is doing really well that we're in the process of rolling out over the next couple of years.

We've always known if we wanted to be able to serve customers the way we think they want to be served and the way we want to, we have to have a mass offering, it starts with perishables. And if you -- if you look at the total spend per household, when we look at the difference between the total amount of spend that we think our consumers and customers are spending on items that we could serve them on versus what they spent in us, the biggest difference is grocery. And a lot of your shops, a lot of grocery starts with perishables. And so we are -- we really are excited that we feel like we found something with shipping perishables out of our same-day facilities.

We've grown our everyday essential is about 3x faster than the rest of our store. It represents about half of our units now if I look at all the everyday essentials. What we find is customers who shop perishables through our same-day delivery, they buy about 7x more everyday essential units than those that don't. So again, if you're trying to close that gap between what consumers spend overall and the items that we can provide versus what they spend with us, perishables has a huge impact on that.

We also see that our grocery customers shop a 4x more non-grocery categories, like home and apparel and a bunch of other categories and spent about 2x more on those categories compared to those who don't use us for grocery. So those are big differences, and we believe that the progress we're making on perishables and grocery is going to ultimately lead to more -- being more relevant to customers for more of their shopping missions.

Our next question comes from the line of Eric Sheridan with Goldman Sachs.

Maybe a 2-parter on market opportunities within AWS. First, with respect to Bedrock, what have been your key learnings about the market opportunity around Bedrock and how it might continue to scale in the years ahead? And then with respect to the decision to sell Trainium externally, how do you think about balancing the market opportunity of external sales relative to the need for compute for that opportunity relative to other opportunities broadly across both internal use cases and AWS?

Yes. On the Bedrock side, I would tell you, you've probably heard this from us multiple times over the last couple of years. But one of the lessons that we've learned I'll say, for 20-plus years in offering technology to customers is that there is not 1 tool to rule the world. People always assume that whatever is in the lead early on is going to be the only thing or things that people use, and it's never the case. It's not true in database. It hasn't matured in analytics. In the beginning of AI, everyone said the only framework people use [indiscernible] much bigger now. And so -- and it's not going to be true with models either.

In -- I'll say in the next 2 to 3 years, I think they're going to be 6 to 8 models that are about as good as each other or at least good enough for 95-plus percent of the workloads. And some will be better reasoning, and some will be better at coding, and some will be better visual pieces and some will be lower latency and some will be lower cost. And most customers are going to use multiple models, and many will use multiple models in the same generative AI application.

And so what customers really want to do, and we see this validated over and over again when we talk to customers is they want a service that has really broad selection of those frontier models and it has very good price performance and it has very strong security and access control and they want that service so that they can kind of decide what combination of models they want to use and the swap them back and forth as they leapfrog each other as different constraints of their applications require them.

And that's what we see with Bedrock. People are very excited about it. We have both closed models and models like Anthropic and OpenAI models. We have open models with Kimi and GLM and others. I think the Western world is very hungry for additional open models that are Western open model. And so I believe that will happen over time as well. But I think that the size of the opportunity in Bedrock is gigantic. And it could be a service as large as EC2, which is a very large service today in AWS. So I think it's very large.

On the balance of chips, and what you sell externally versus what you use inside your cloud. It's a very good question, and it's a question that we're continuing to evaluate as we speak. I would say that the number of companies who have a significant cloud business and a chip strong enough to sell externally is effectively too. So there really aren't many customers that have to grapple with the same type of problem at this type of scale.

And at this point, everybody is producing their advanced chips with the same provider. There's only so many chips that provider can provide each year, everybody gets an allocation. And you have to decide with that allocation, how much of it you want to sell externally and how much of it you want to use in your cloud. And there's a trade-off. If you sell externally, you're able to realize more revenue in the short term because you get to recognize the revenue as soon as it leaves your inventory.

And on the flip side, if you decide to use it inside your cloud in the form of EC2 instances or Bedrock, you actually will end up driving more absolute revenue and free cash flow from those chips because you get to use them over 5 to 6 years and drive revenue for a longer period of time. So -- it is -- you have to really think through what you want to do. To date, I would say that we've taken a different approach in how we've used our chips than the other provider who has the same type of balancing act going on.

And as I mentioned earlier, we believe we're going to have a very significant business selling chips externally and we're going to start in late 2027. And we will be thinking through that balance, I think probably every year. They'll both be very significant. Our ability to create price performance scalable chips is going to be a very significant strategic advantage for us at scale, both in our own cost structure as well as in building a separate business and selling external chips. And selling internal chips really is a very different business than your cloud infrastructure business.

Our next question comes from the line of Doug Anmuth with JPMorgan.

When you think about the 40% plus AWS revenue growth each of the next 2 years and then the multiyear path beyond that, can you just talk about the CapEx required to drive that kind of revenue over the next couple of years and beyond. And then also, Andy, any updates just to how you're thinking about the duration of returns on both servers and networking and then also data centers?

Yes. I'll take them in reverse order. So on the breakeven on our infrastructure. As I mentioned in my opening comments, we're continuing to move that left. We're continuing to make that shorter. There are a lot of -- core is a little bit different from AI. It's a little bit different from how you finance these things. And so it varies depending on which you're talking about, but we're continuing to make progress.

As we historically have been pretty good at doing, if you think about AWS at finding ways to extend useful lives and to make the breakeven sooner. And we're making good progress there, too.

On the -- on the CapEx involved to grow the way we're talking about, it's meaningful. And I think that this -- the way we think about it, Doug, is that this is a once-in-a-lifetime opportunity. I think what we're seeing with AI is the biggest transformation in our lifetime. And if the idea of being able to grow 39% with $184 billion annual run rate, and to grow 40% a year on top of that in each of the next 2 years and ultimately end up as a $1 trillion plus business, we're going to invest to continue being the leader. We are the leader today in the, I'll call it, the infrastructure and AI space.

And would it be more convenient if this type of acceleration and opportunity happened over 25 years and you could spread the CapEx over 25 years, yes, it might be more convenient. But that's not the world that we live in, and that's not the opportunity. And we can fund this plan. And I think it will forever change the size of Amazon -- will ever change -- forever change our revenue, it will forever change our free cash flow. And we're going to invest and we're going to pursue it to continue to be the significant leader like we are today.

And our final question comes from the line of Ross Sandler with Barclays.

Yes, I thought I'd double down on the Bedrock question. So the token marketplace on Bedrock seems to be one of these the hottest services in AI right now. And our understanding is the majority of these workloads are running on Trainium. So as this becomes a bigger part of the overall AWS story like you just mentioned, Andy, how might that impact both revenue growth, but especially margins at AWS as you kind of shift from selling bare metal to selling more inference as a Service. Could you just talk about that?

Yes. Well, you're right that the majority and most of Bedrock today is being run on top of Trainium. And we do that in meaningful part because we really like the price performance that we're getting, and it allows us to keep -- to both keep prices as low as possible for customers as well as our ability -- we like the margins associated with it. If you think about margins in the cloud space and you're looking at different players and what their cost structure looks like, there are a few things to pay attention to.

The first is for their data centers, how many of them are they actually buying or long leasing versus doing colo or neoclouds. I mean nelclouds are about 3,500 basis points worse in operating margin than if you long lease them. So -- and then you look at what is the hardware on top of which they're running. And the -- if you're using the hardware that everybody else is using, you'll have a comparable cost structure to what everybody is doing. And if you have the ability to use hardware that's different, you'll have an advantage.

And so I think when you -- when you look at what we're doing, most of our capacity is either owned or long leased, which is -- enjoys better operating margins than what you find when people are mostly doing colo or neo or a big chunk in color or Neo. And the fact that we're running on top of our own chip that's providing both advantaged price performance and is more cost effective for ourselves because we own it, I think, gives us some structural cost advantages that I think are going to be very helpful to customers as we pass them back to customers, and helpful to the overall economics of AWS.

Thanks for joining us for the call today and for your questions. A replay will be available on our Investor Relations website for at least 3 months. We appreciate your interest in Amazon and look forward to speaking with you again next quarter.

Thank you for standing by. Good day, everyone, and welcome to the Amazon.com Quarter 2 2026 Financial Results Teleconference. [Operator Instructions] Today's call is being recorded.

And for opening remarks, I will be turning the call over to the Vice President of Investor Relations, Mr. Dave Fildes. Thank you, sir.

Hello, and welcome to our Q2 2026 Financial Results Conference Call. Joining us today to answer your questions is Andy Jassy, our CEO; and Brian Olsavsky, our CFO. As you listen to today's conference call, we encourage you to have our press release in front of you, which includes our financial results as well as metrics and commentary on the quarter.

Please note, unless otherwise stated, all comparisons in this call will be against our results for the comparable period of 2025. Our comments and responses to your questions reflect management's views as of today, July 30, 2026 only, and will include forward-looking statements. Actual results may differ materially.

Additional information about factors that could potentially impact our financial results is included in today's press release and our filings with the SEC, including our most recent annual report on Form 10-K and subsequent filings.

During this call, we may discuss certain non-GAAP financial measures. In our press release, slides accompanying this webcast and our filings with the SEC, each of which is posted on our IR website, you will find additional disclosures regarding these non-GAAP measures, including reconciliations of these measures with comparable GAAP measures. Our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our results are inherently unpredictable and may be materially affected by many factors, including fluctuations in foreign exchange rates and energy prices, changes in global economic and geopolitical conditions, tariff and trade policies resource and supply volatility, including for memory chips and customer demand and spending, including the impact of recessionary fears, inflation, interest rates, regional labor market constraints, oral events, the rate of growth of the Internet, online commerce cloud services and new and emerging technologies and the various factors detailed in our filings with the SEC.

Our guidance assumes, among other things, that we don't conclude any additional business acquisitions, restructurings or legal settlements. It's not possible to accurately predict demand for our goods and services, and therefore, our actual results could differ materially from our guidance. And now I'll turn the call over to Andy.

Thanks, Dave. We're reporting $200.6 billion in revenue, up 20% year-over-year. Operating income was $27.5 billion, up 43% year-over-year. Q2 was another very strong quarter for Amazon. I'll start with AWS, which is booming right now, and I'll share the numbers what we think is going on and why we're enthusiastic about the ROIC equation, even with heavy CapEx the next few years.

First, the numbers. Revenue growth of 36.7% year-over-year, accelerating for the fifth straight quarter, our fastest growth in 18 quarters back when AWS was less than half of its current revenue size. We added over $4.6 billion in revenue quarter-over-quarter, about 80% more than our largest increase ever, and our backlog stands at $496 billion, growing triple digits year-over-year.

AWS is now $169 billion annualized revenue run rate business, which for perspective, we placed at 24th on the Fortune 500 list if it was a stand-alone company. Our chips business now has an annual revenue run rate of over $25 billion, growing triple-digit percentages year-over-year. Our AI revenue run rate climbed significantly quarter-over-quarter has now also over $25 billion, growing triple-digit percentages year-over-year. Customers choose AWS because we offer the broadest capabilities, they want their AI inference to reside near their other applications and data and more of it resides in AWS than anywhere else and because AWS has the strongest security and operational performance.

We're seeing strong growth across both AI and non-AI, what we call core and growth in 1 is driving growth in the other. Growth in AI drives core because post training reinforcement learning and agent tool use is mostly done on CPUs versus AI accelerators. This is an advantage for AWS as our Graviton chip is the strongest CPU chip hovering up to 30% to 40% better price performance than other options. You need a place to store the AI data and to run vector databases which are also emblematic of a meaningful edge for AWS because we have the broadest and most capable functionality by a fair bit in these core infrastructure areas.

We feel similarly about the AI stack top to bottom. We have a unique offering that customers are excited about. As we've been saying for 18 months now, technically competent companies are going to build their own foundation models, not the really big frontier models, but smaller models that leverage their proprietary data. There is no easier service for this than our SageMaker AI service. Customers also need a high-performance, cost-effective inference service, and that's what Amazon Bedrock provides. Bedrock not only provides the best selection of leading models as superior performance and with the governance and security controls the companies need, it's also continuing to grow incredibly quickly.

In addition to leading model building and inference services, customers need easier ways to build, run and leverage agents. For example, even if you've built an agent, you have a lot of mock to worry about. A production agent needs somewhere secure to run, memory, so it holds context, and identity, so it can act on a user's behalf, tools and data to connect to and a way to watch what is doing once real traffic hits. Stitching all that together reliably is hard and it stalled many production deployments. It's why we built Bedrock Agent Corp. to provide [ Box's ] managed infrastructure, and our teams keep iterating recently adding features like policies, which give companies deterministic controls over what agents can do, payments so agents can execute transactions autonomously, web search to ground agents knowledge without having to leave AWS and a new harness that further speeds up how fast customers can put this all together, including creating the agent with strands. .

While companies will construct their own purpose-built agents from the ground up, most will also use turnkey agentic services. Coding agents are a good example and there are several successful ones, including Claude Code, Codex and our own spec-driven Kiro, which is up to 50% more cost effective than others and tripled in usage quarter-over-quarter.

Another of these agentic services is Amazon Quick, an intelligent AI work companion that helps you manage, search and automate your digital workload across e-mail, calendar, local or cloud files and custom workflows. Unlike other offerings in this space, Quick also lets you manage across leading SaaS tools like Slack, Salesforce, Jira, Teams and ServiceNow. Quick enforces the company's existing access control, so each person sees only what they're clear to see, and then it takes action, scheduling meetings, drafting and sending e-mail, updating a CRM record, building a dashboard and more.

In Q2, we made Quick even more capable adding autonomous agents that customer set up and plain language to run continuously in the background and carry out multistep tasks, a personalized activity feed that pulls e-mail, messages, calendars and tasks into 1 prioritized view and 16 new integrations, including Adobe, [ Moody's ] and Snowflake. Quick is momentum with 3M, Allianz, AstraZeneca, Autodesk, BMW, Exxon, FINRA, Hyundai, Intuit, Mondelez, Moody's, the NBA, the NFL life insurance and Southwest Airlines all using it.

We also have services like Amazon Connect, our call center service and AWS Transform, which automates software migration growing quickly. I'll mention 1 more that I'm excited about as frontier models get increasingly powerful, they're making it easier to find security vulnerabilities and technical applications, many of which humans haven't found. This is obviously concerning for companies that protect important data.

We recently released AWS Continuum, which discovers, prioritizes, validates and remediate code vulnerabilities. It starts by ingesting the backlog of vulnerabilities the team already has and then leverages the new frontier models to run comprehensive scans. Continuum uses agents in each company's own business context to prioritize what matters. Reasoning through questions like, is the effective component deployed? Is it reachable? Is it in a production path? And what's the impact if it's exploited and then validate vulnerabilities in the sandbox, so teams aren't chasing false positives, and finally, it recommends the fix. It's hard to talk with enterprises about AI right now without their mentioning security, and we expect Continuum to grow quickly.

I mentioned earlier that our chips revenue run rate is now over $25 billion. We're unusually well positioned for this AI inflection given our leading price performance chips in both AI with [indiscernible] and CPU with Graviton. In addition to the 2 leading AI labs in the world in Anthropic and OpenAI making multiyear multi-gigawatt commitments to Trainium, an increasing number of AI startups are also adopting Trainium including unicorns like Neurorobotics and Odyssey, joining start-ups like 12Labs, Descartes, Poolside, Karakuri, Medigene, NetoAI and Splash Music in larger companies like Uber and Pinterest, all adopting Trainium.

Graviton is used by 98% of our top 1,000 EC2 customers. The revenue commitments have increased nearly 3x quarter-over-quarter and Graviton 5 is growing nearly 2x faster as Graviton 4 did. We also continue to have a deep partnership with NVIDIA, and we'll continue making AWS the best place to run NVIDIA chips as we have customers who will run on NVIDIA for as long as we can foresee, and we believe strongly that customers want choice. Choice is good for customers, competition and driving the cost of inference now, which customers care deeply about.

Let me talk for a second about how we see this investment playing out. Earlier this year, we said we plan to invest approximately $200 billion in cash CapEx in 2026, the majority of which to support AI and AWS. And this level of spend and higher, we have clear line of sight to strong financial returns. I'll explain why. There are 2 major parts of the investment, the data centers and the servers and networking equipment that go into them. These are different capital cycles.

Data center capital is spent starting 2 years before we can put servers into them to start monetizing. Once a data center opens with servers plugged in, we start generating significant revenue right away and then get to monetize these data centers for 30-plus years without having to spend that start-up capital again.

Servers and networking equipment operate on a shorter cycle. We typically purchase these a few months before putting them into service, so we have strong visibility into customer demand before we trigger the spend. If the demand isn't there, we won't spend the capital. For servers and networking equipment, on average, it takes a little less than 3 years to break even on that investment. The servers currently have a useful life of at least 5 to 6 years. And most of our AI capacity these days is being contracted for at least 5-year terms. That means that we're driving significant free cash flow on the servers and networking equipment in the 2 to 3 years after we break even.

It's also worth noting that AWS has a strong track record of pulling forward breakevens on server equipment, where we've already made meaningful progress and finding ways to extend the useful life of this equipment without sacrificing customer experience. So for our data centers, which have 30-plus year useful lives, we should get at least 5 to 6 generations of server economics, like I explained earlier, with subsequent generations after the first having even better overall economics because we don't have to repeat that upfront data center investment I mentioned earlier. This means in the short term, when demand is necessitating so many data centers being built simultaneously in advance of when we can start monetizing them. We'll spend a lot of CapEx and encounter free cash flow headwinds and until these data centers come online can be monetized and we get a few years into these servers being utilized. But as we get a few years out, the revenue growth outpaces the incremental CapEx growth, which will happen at some point. The resulting revenue, free cash flow and return on invested capital is very compelling.

We've done this before in the first era of cloud computing just over a longer time horizon where demand built more gradually than it has in AI but we see the margins and returns in AI tracking what we saw with core at the same point of evolution actually a little ahead. We now believe we will spend approximately $220 billion in cash CapEx in 2026. The higher cost of memory pushing this number up from a prior estimate of about $200 billion. But even at that amount, we will still not have enough capacity to meet all the demand we have in 2026. And I believe this dynamic will also be true in 2027 too.

In fact, the demand we already have for 2028 is striking. And remember, enterprises are still very early in using inference at scale in their current production applications. We won't believe AWS could become a few hundred billion dollar revenue business and now believe it will be at least double that, and very positively be $1 trillion annual revenue business for us in time with very appealing accompanying free cash flow and return on invested capital.

I'll now turn to stores. We added millions of new products to our selection, including over $700,000 from notable brands. We also expanded ultra low price selection on Amazon Haul in the U.S. by nearly 20x since launch, and now have over 6 million items priced under $10. We continue offering everyday low prices to met or beat our competitors as well as deep discounts and savings during sales events.

We're pleased with the customer response to Prime Day where customers shop millions of deals, including more than 80% at our lowest price of the year and hundreds of thousands discounted by 40% or more. We're the second largest grocer in the U.S. and our grocery business continues growing quickly across perishables and nonperishables. The number of monthly active perishables customers grew over 50% since the start of the year same-day orders with perishables average over 3x more units per order and fresh groceries now make up 6 of the top 20 best sellers in Amazon.com.

We grew the number of new customers for Amazon Pharmacy by more than 2x in the first 6 months of the year and same-day prescription deliveries nearly 5x. We also saved customers nearly $250 million so far this year in out-of-pocket costs, up more than 400% year-over-year. We continue speeding up overall delivery and once again achieved record delivery speeds for customers in the first half of the year.

We offer millions of items for same-day delivery with Prime, up to 40x more selection than a typical big box retail store, and our same-day network continues to expand. Globally, we delivered over 40% more items same-day or overnight in the first 6 months of the year than the same period last year. We also continued to expand our ultrafast service Amazon Now, which offers delivery in 30 minutes or less on thousands of everyday essentials. We added 80 new cities and towns across the U.S. and several major cities in Egypt in Q2. And Amazon now is available in 9 countries and over 250 cities and towns globally.

We continue to see strong customer response with over 80% growth in gross sales and units sold quarter-over-quarter, and we served over 60% more customers quarter-over-quarter.

We recently launched Amazon Supply Chain Services, so any business can move, store and deliver everything from raw materials to finished products using the same supply chain that supports Amazon, and we already have several large customers, including Procter & Gamble, 3M, Plans and American Eagle Outfitters. The stores team also continues to innovate and deliver for customers with AI. Customers love Alexa for shopping or agentic AI shopping assistant. It offers personalized recommendations, product comparisons, price history and the ability to automate shopping through features like price rewards and auto buy.

Over 350 million customers have used it in the last 12 months, an engagement accelerated in Q2 with active users nearly doubling and interactions up over 5x year-over-year. We also expanded Amazon Lens, which lets customers take a photo of anything they see and instantly find the same or similar items on Amazon to 10 additional countries is now available in 21 countries around the world.

Moving on to Amazon Ads. We saw strong growth across our offerings, generating $19.8 billion of revenue, up 26% year-over-year. Sponsored Products continues to be our largest offering and a key driver of growth. Additionally, increasingly more shoppers are discovering products in our agentic and conversational experiences, including an Alexa Plus and Alexa for shopping. Shoppers who click a sponsored prompt convert to a sale of 48% more often and spend 21% more on average than those who don't.

We see continued growth and engagement in Prime Video ads and live sports. We introduced more than 30 new advertisers to the NBA in our first year, and inventory on Thursday night football, NBA, WNBA and NASCAR all sold out. Advertisers are increasingly investing in multisport strategies with brands activating across multiple sports, seeing 2.3x higher unduplicated reach compared to single support advertisers and multisport viewers are driving 12% higher spend and 17% more orders on Amazon.

And finally, we make it easy to create, launch and optimize full funnel campaigns using AI-powered tools including Ads Agent, which turns hours of setup and targeting into minutes. Advertisers using as ads agent targeting see 8% lower cost per impression and 6% lower cost per acquisition, and we've expanded it to 11 new countries this year.

We're also continuing to see momentum in several other areas, and I'll mention just a few. Starting with entertainment, the inaugural season of NBA on Prime Video delivered strong viewership with a peak of 6.5 million U.S. viewers for Game 7 of the Eastern Conference semi files, outperforming Game 7 on broadcast a year ago. In Europe, viewership in the MBA more than doubled year-over-year on Prime Video, with the highest average viewership on record. We also drew 36 million viewers globally for the Series Premier of off-campus on Prime Video in its first 12 days, becoming Prime Videos #3 top viewed series debut ever. [ Solexa ] plus expanded to Germany, Austria, France and Brazil and hundreds of millions of customers are using new Alexa experiences.

We find that everywhere less it goes, it drives momentum for the business. For example, in the U.S., customers who use Alexa for shopping spend an average of over 40% more per order than those who don't. And customers who've tried Alexa Plus are signing up for Prime at nearly 25% higher rates. And finally, Amazon Leo is close to 400 satellites in orbit, enough to begin initial satellite Internet service this year. We already have meaningful revenue commitments from enterprise and government customers and we have more than 20 partners who will extend the reach of our network across the globe.

We continue to be in the middle of some of the biggest inflections of our lifetime and we're building multiple new long-term businesses that will make customers' lives better and easier and lead to substantial free cash flow and return invested capital for our shareholders and business.

With that, I'll turn it over to Brian.

Thanks, Andy. Starting with our top line financial results. Worldwide revenue was $200.6 billion, a 20% increase year-over-year, excluding the impact of foreign exchange. These results include the timing shift of Prime Day into Q2 for most of our largest countries, including the United States. In Q2, we reported worldwide operating income of $27.5 billion. This includes the benefit from 2 items that reduced expenses by approximately $1.2 billion during the quarter.

First, we received tariff-related refunds of approximately $600 million. This is included in our North America segment and represents a significant majority of refunds we expect to receive. Second, we recorded a separate benefit of approximately $600 million related to the change in fair value measurement of energy contracts subject to derivative accounting. This primarily impacts the AWS segment. These energy contracts are to secure electricity supply for existing and future operations.

Certain terms in these contracts cause them to be subject to derivative accounting. Derivatives are measured at fair value each reporting period and changes in fair value measurements will create unrealized gains and losses recorded within operating expenses on our statements of operations. While the impact of these fair value measurements can vary, these adjustments have not been significant in prior quarters.

Moving on to our segment results. In the North America segment, second quarter revenue was $116.2 billion, an increase of 16% year-over-year. International segment revenue was $42.2 billion an increase of 15% year-over-year, excluding the impact of foreign exchange. Worldwide paid units grew 17% year-over-year. Prime remains a key pillar of our business and our double-digit year-over-year membership growth reflects the value of our offerings delivered at scale. We saw broad-based momentum across the inputs that drive our customer experience, including new selection, sharp prices and fast delivery.

Shifting to profitability. North America segment operating income was $9.1 billion, with an operating margin of 7.9%. International segment operating income was $1.7 billion with an operating margin of 4.1%. In our fulfillment network, we made progress optimizing inventory placement, shortening shipping distances, reducing touches per package and improving consolidation rates. We're expanding our deployment of robotics and automation, which have been integral to our operations for decades. We're retrofitting our facilities with our latest generation technology, and we expect to more than double our fleet of robotic arms like [ Cardinal and Spero ] in 2026.

We continue to lower our overall cost to serve even as we faced heightened transportation costs driven by fuel inflation from the conflict in the Middle East and higher linehaul rates from driver capacity limitations. Excluding the impact of higher fuel and linehaul rates, shipping costs grew more slowly than worldwide unit growth at a pace that is relatively consistent with last quarter.

On the operating income side, the impact of these higher costs is partially offset by our FBA fuel and logistics surcharge that was implemented in April. Looking ahead, we see meaningful opportunities to further enhance productivity across our global fulfillment network, all while continuing to raise the bar and delivery speed.

While our operating margin may fluctuate and the progress may not always be linear, we take a deliberate approach to achieving sustained long-term improvement in our cost to serve.

Moving to the AWS segment. Revenue was $42.2 billion, up 36.7% year-over-year, driven by both core and AI services. AWS now has an annualized revenue run rate of $169 billion. Customers continue to increase cloud migrations and scale up their use of AWS core services. And increasingly, customers seeking the full benefits of AI are accelerating their transition to the cloud. We see a strong linkage between AI spend and core growth. As customers invest in AI, we see a core spending increase in core consumption.

We expect this relationship to strengthen over time as more AI workloads move into full-scale production and drive additional demand for our core services. AWS operating income was $16.6 billion which reflects our strong growth, coupled with our focus on driving efficiencies across the business. Our investments in software and process improvements, optimize server capacity and help to develop a more efficient network using our lower-cost custom silicon and custom network gear.

Now turning to our cash CapEx, which is $53.1 billion in Q2. This primarily relates to AWS and generative AI as we invest to support strong customer demand. We'll continue to make significant investments, especially in generative AI as we believe it to be a massive opportunity with the potential to drive long-term revenue and free cash flow.

Before moving on to guidance, I'll briefly touch on the impact from the tariff-related refunds on our results. We are participating in the tariff refund process. And as I mentioned earlier, we received approximately $600 million in Q2. The amount is limited for a couple of reasons. First, our teams did a lot of work forward buying and prepositioning inventory to avoid tariff costs. Second, we are not the importer of record for the large majority of items sold in our store given suppliers typically handle imports and pay relevant tariffs. In cases where we did see an increase in costs due to tariffs, we largely absorbed these costs rather than pass them on to customers.

You can see that in how we stayed very sharp in our price throughout the last year with our product prices on average 14% less than other retailers according to third-party research firm, Profitero. We have identified a limited set of circumstances where we can trace that we pass specific import charges on to customers. And when we receive those refunds, we will proactively contact effective customers and automatically issue refunds to them. Otherwise, like other large retailers will utilize refunds to continue to invest in low prices for customers.

I'll finish with our financial guidance. Q3 net sales are expected to be between $197 billion and $202 billion. I'll call out 2 areas that are driving the sequential deceleration in net sales growth from Q2 to Q3. First, Prime Day timing shifted this year with the sales event occurring in Q2 for most of our large countries, including the U.S. In 2025, Prime Day was entirely in Q3. Excluding the impact of Prime Day in both 2025 and 2026, third quarter 2026 year-over-year growth would have been nearly 400 basis points higher. We recognize the timing shift of Prime Day creates some noise in the modeling of our financial results. However, after adjusting for Prime Day, as we look across our stores business heading into Q3, we see strong customer engagement growth and a continuation of the trends that we have seen in the first half of the year.

Second, the Q3 guidance anticipates an unfavorable impact of approximately 80 basis points from the year-over-year changes in foreign exchange rates based on current rates. Q3 operating income is expected to be between $22.5 billion and $26.5 billion. I want to thank our teams across the company for their hard work and focus on customers. We'll continue to work to deliver more value to the broad range of customers we serve across each of our businesses, which is the only reliable way to create lasting value for our shareholders.

With that, let's move on to your questions.

[Operator Instructions]And the first question comes from the line of Doug Anmuth with JPMorgan.

Just 1 for Brian and 1 for Andy. Brian, many have assumed that AI workloads would be lower margin, at least near term. Can you just talk about the drivers of the 39% AWS operating margin in 2Q and just how we should think about sustainability? And then, Andy, strong bedrock traction with customers spending more in the quarter than in all the prior quarters combined. But when you think about the full stack offering. Does Amazon need its own leading model towards the frontier?

Doug, let me start with the first question. We're pleased with the growth in both revenue and also margin expansion that we had in AWS in Q2, especially given the size of our business. You're seeing, despite the large investments, AWS margins have continued to remain strong, and we're up 650 basis points year-over-year. 520 basis points if you exclude the derivative accounting gain that I mentioned.

We've said before these margins will fluctuate, they're based on a number of factors, including our investment levels, Alexa products, Alexa AI versus non-AI, but I would say that the profitability you're seeing from AWS isn't random, it's a result of disciplined efficiency gains, capacity optimization, which we've benefited quite a bit from in Q2 and always closely managing our fixed costs. So again, they'll fluctuate, but very strong performance year-over-year, and we'll take it.

Yes. I'll just add 1 other thing to what Brian said, which is, as I mentioned in my opening comments, we see the AI business following very much the same type of margin trajectory that we saw in the core business before and it's a little bit ahead of that pace that we saw. So we're optimistic about that.

On the question about Bedrock and our own Frontier model. So my view of it is that AWS and Amazon can have a wildly successful business without its own Frontier model. And a lot of that is because there is not going to be 1 model to rule the world. You already see that right now. You see it. It's not just Anthropic or it's not just OpenAI. You see increasingly more and more companies being interested in the open models as well. And we have all of them in Bedrock.

And it's 1 of the many reasons why Bedrock is growing so quickly. If you're a company that's building important AI applications, you want to make sure that you have the ability to use all the available models. They're going to each leapfrog each other at different times. They're going to have lots of different models that actually are comparable in capabilities. And you want that leading selection with the right price performance and with the right governance and security and there's nothing like Bedrock that provides that right now. And we use those models as well.

All that said, we are pursuing our own Frontier model. And we're doing it for a few reasons. First of which is it just gives us additional control over cost, cost for our own consumer applications, but also we're trying to drive costs down for customers. And having a player like ourselves, it's always focused on trying to take the price performance and the cost down for customers all the time, we think will help keep the models more cost effective for customers.

I think also it allows us to have more control over prioritization of what models focus on. And we have, both from our own external customers as well as our internal customers inside the company, certain priorities that matter that we want the models trained especially well for. And then it gives us some control on speed. And so my view of it is that within the next few years, you're going to have at least a half dozen models that are comparably good to each other. They'll all be in Bedrock and 1 of them will be ours.

The next question comes from the line of Justin Post with Bank of America.

Just thinking about the AWS acceleration, was that really driven by a lot of capacity coming online in the quarter. You guys have been very more open than your peers on gigawatts you're adding. Any help you can give us on how much you might be adding in the second half versus the first half and also how you're thinking about '27.

Yes. I think that -- there are several reasons for why we saw such significant growth. And it's -- we're really pleased and excited fifth straight acceleration quarter, largest acceleration in 18 quarters. I think there are several things that are driving it. I think the first part is that customers are choosing AWS in part because it has the broadest functionality across both cloud core and AI in part because it has the strongest operational performance in security and in part because as more and more companies are bringing their inference workloads to production, they wanted to live near the rest of their workloads and data and so much more of it lives in AWS than anywhere else.

I think there are a couple of other things going on here on the core side. I mean, AI is obviously growing at a very rapid rate as we talked about well over $25 billion in annual revenue run rate at this point. But the core business is growing very quickly as well. And I think there's at least a couple of things at [ pipe ]. One is that increasingly more enterprises are building their transformation plan to move from on-premises to the cloud.

Remember, by the way, that 85% of the global IT spend is still on premises. That equation is going to flip in the next 10 to 20 years. And you see more and more enterprises that are moving and building plans to move to the cloud, and we're winning the lion's share of those with the capabilities I mentioned earlier and the advantages.

And then AI is growing at such a rapid rate, and it's pulling a long core alongside of it. And that's because the post-training and the reinforcement learning and all the agent tool use is being driven on CPU and core. And with the leading CPU chip and Graviton it makes AWS an even more attractive choice.

So yes, we're adding a lot of capacity, but there are a lot of other reasons why it's growing. I expect that we will -- we're on pace with the capacity build that we talked about a few quarters ago where we said we expect to have double the capacity -- power capacity by the end of '27 that we had in '25, and we continue to be on that track.

The next question comes from the line of Brian Nowak with Morgan Stanley.

I have 2, Andy. I appreciate the color on the long-lived data center investments versus the server and network investments. Just the question is, as you sort of look into 2027, you look at the demand that's coming, et cetera. Are you at a point where you're going to be able to start to slow that long-lived data center spend at all in '27? Or is that just too soon where you're still going to have to be opening up new data centers over the next 2, 3, 4 years as you look into '27 that's number one.

Second one, in the past 90 days or so, you've talked -- the company talked publicly about selling Trainium at some point to third-party data centers. How do you think about when you could do that and just sort of the ROIC on that versus core AWS loads?

Well, on the first question, Brian, we have so much demand right now. Apart from what we've talked about in '26, a lion's share of capacity in '27, we're adding a lot of capacity, as I mentioned just a few minutes ago, is largely reserved. And we have quite a bit of capacity that's already been reserved for '28. And so I think it's actually kind of useful to look at least our view of what we see in the demand and adoption curve right now, which is we see this adoption curve in AI right now is very barbell. There is on 1 end of the barbell the AI labs are consuming gobs and gobs of compute, and there are a few runaway successful generative AI applications like Claude Code and ChatGPT.

On the other end of the barbell are enterprises who are getting real value from AI in cost avoidance and productivity. And these are things like automating customer service or business process automation or fraud or things like that. In the middle of the barbell, is all of the current enterprise production workloads. Some of which have -- are using inference in a pervasive way, but most of which aren't. And that is going to change very significantly over time. And that will be, in my opinion, that will be the largest absolute segment, the existing production workloads in the enterprise and new businesses and workloads that start-ups build too. And so I think we're still in the relative early stages of how much demand there's going to be for AI.

I think it's going to change every customer experience that we know I think that it will invent all sorts of new ones that we never imagined. And I don't know if the trajectory of that middle part of the barbell will be the same wildly steep trajectory that we've seen with the current barbelled AI Labs piece. But we have a lot of demand in front of us, and we're going to invest in this business to continue to be the significant market segment leader that we are today.

We think, as I mentioned earlier, it has the potential to be a $1 trillion revenue business for AWS, and we intend on continuing to be the leaders.

On the question about selling Trainium, we're quite excited about what's happening in our chips business. As I mentioned earlier, it's over $25 billion in annual revenue at this point. We think we have the leading price performance chip in both the AI space with Trainium and in the CPU space with Graviton. In fact that we have a multiyear multi-gigawatt commitments from the 2 largest AI labs in Anthropic and OpenAI and more and more companies, as I mentioned in my opening comments, using Trainium is exciting and promising.

And we have -- we just have an incredible amount of demand for Trainium. And so there are a lot of customers who are very excited about using it in the form that we're providing right now. We do have an increasing number of customers who are interested in us providing the Trainium chips to them separate from even from our cloud and we're actively having those conversations and exploring, and I expect there's a real chance we'll do that in the future.

The next question comes from the line of Colin Sebastian with Baird.

Andy, is it fair to say that there's a more concerted effort to move into the application layer with Kiro and transform. And I guess more broadly plans for workplace productivity tools. And do you see those as sort of providing a boost to the broader platform offering from infrastructure on up?

and then, Brian, I mean, just given the demand signals you guys are both talking about and plans for additional capacity, what are your current thoughts on sources of capital for the build-out over the next couple of years?

Colin, let me take your second question first. You've seen us issue debt this year. We have a lot of options available to us as we continue to fund this growth that we're seeing in AWS. So we'll continue to look at all the options and make the appropriate decision at the right time, but nothing to share today. .

And your first question, Collin, we see we see a very substantial opportunity, both for our customers and for AWS in building some of these Agentic applications. And some of this is borne out of what customers tell us, they wish they had and they want to be using. Some of it is borne out of just needing to provide those capabilities to ourselves inside Amazon. And so Kiro, which is our coding -- agentic coding service is an example of that. But Amazon Quick is a really interesting example where we just had so many people inside the company who wanted really an intelligent AI assistant to help them work to even first started off with -- we're a very document-oriented, culture. People wanted when they got documents not to have to read every document so carefully to get a summary and then to be able to write their own analyses and responses to these things and to be able to do business intelligence through these agents.

And that's really how Amazon Quick started was to do research, to do business intelligence, to do summarization. And then we had so many people inside the company using it that they said, can't you actually find a way to make it much more productive and easier for us to manage our e-mail, to manage our Slack communications, to manage our calendar and to use all those things together. And that's really this next instantiation of Amazon Quick.

And as I mentioned in my opening comments, it's pretty remarkable, not only how fast it's taken off inside Amazon, but how many external enterprises have put it into production with a very large number of people at their companies. We see that opportunity up and down kind of the different needs of companies. I think that Connect, Amazon Connect, which is our call center service, which is used by all 5 major leading airline providers as well as many of the leading banks and health care companies continues to grow very quickly.

AWS Transform, which makes it much easier to migrate software is super useful for enterprises. And then the latest 1 we just launched with Continuum. It's really hard to have a conversation with a large company about AI right now where they don't actually ask you about security with just all the noise and the and the hype about the security risk with the most current powerful models. And so Continuum really allows them to use those models productively to find their own vulnerabilities in their code to design the fixes and to help them deploy them.

And those are kind of the first set of them. They all have very high promise, but there are several others that we're working on, and we think that's going to be very helpful for customers and our business.

The next question comes from the line of Jason Holstein with Oppenheimer & Company. Okay. The next question comes from the line of Ken Gawrelski with Wells Fargo.

Two if I may, first, your RPO reported is 2.5x that of the third quarter of '25 when you gave us the doubling of capacity comments for year-end '27. How does that RPO number and the massive expansion there? Impact your outlook for future capacity? I know that you talked about through year-end '27, but maybe at least [indiscernible], if you could touch upon what the RPO means for '28 and beyond capacity.

And the second 1 related is you raised your CapEx guidance this year for some supply chain inflation. Could you talk about how your pricing strategy at AWS incorporates future cost inflation. Your longer-term contracts allow for stable return profiles despite cost inflation?

Yes. I'll start on the backlog number, yes, to your point, it's very substantially continuing to grow. I think it's a reflection, again, of customers being very enthusiastic about using AWS, both across core and for AI. We have taken in -- we know about that backlog, obviously. So that's all taken into account in our CapEx projections. And over time, I expect that we will continue to sign more deals with customers. And as I mentioned earlier, we're going to pursue the opportunity to continue being the significant market segment leader that we are.

On the second question on the supply chain inflation, what I would say is that most of the deals that you sign, other than -- there's a certain amount of your demand that is on demand, where there aren't contracts. But a large amount of it tends to be deals and agreements that you've signed. And the deals that you signed, those will be the prices and those would be the agreements that we have over the duration of that contract and new agreements that you signed, you always take into account what your costs are and how you ultimately build a price that you agreed to with your customers.

And I think it's no secret right now to any company in the world that there are inflated prices right now on some of the components like memory and hard drives and SSDs.

Our final question comes from the line of Eric Sheridan with Goldman Sachs.

Maybe pivoting to the Commerce business, when you're talking about scaling some of your initiatives around fast commerce and a wider array of supply of groceries and everyday essentials. Can you talk a little bit about the signal you're getting from consumers in terms of either adoption rate of those services or what it's doing the overall spend trends? And whether there's any countries or geographies where you're seeing different outcomes as you launch and scale some of those services.

Yes. We are quite enthusiastic and excited about the pace with which we are growing the amount of everyday essentials as well as perishables in the business right now. And I think some of that has to do with the broader selection we have. And some of that has to do with just how fast our delivery has gotten over the last 2 to 3 years. And when you can deliver items to people, when you have that broad selection like we do at low prices and you can deliver that selection to people as fast as we are right now, people consider you for a lot more of their total purchases and shopping visits.

And we've talked for a while about grocery. And we have a very large business in that space. It's -- last year, it was over $150 billion in gross merchandise sales, making us the second largest grocer in the U.S. And a lot of that are nonperishables. The middle aisles that you'd find in a grocery store where you have consumables and canned goods and beauty items and pharmaceutical items. A big chunk of it is our Whole Foods market business, which is the leading organic grocer out there.

And if you look at the growth in geographies that Whole Foods markets are in, they're significantly outpacing the growth of comparable grocers and the profit trajectory continues to trend the right way as well. And we found a new format there in daily shop in urban settings, it's off to an amazing start that we're expanding very rapidly.

And then we always knew that if we wanted to serve the number of customers who want us to serve them and that we want to serve, we had to find a way to offer mass brands and perishables in a significant way. And we've tried lots of experiments over the last few years, as we've talked about on this call for a few years, but we have finally found something that is a real needle mover for us in offering perishables in our same-day facilities that we're now able to offer same-day perishables in 2,300 cities around the U.S.

If you look in those cities, 9 of the top 10 best sellers in those geographies are perishables. The number of monthly active perishable customers has increased 50% since the start of this year and for same-day orders with perishables, on average, the average 3x more units per order, and so we're just seeing very significant traction in our everyday essentials and in our grocery items. And we're not done experimenting, by the way, with other physical formats in the grocery side, but we've hit on something with same-day perishables in our same-day facilities as changing the trajectory of our everyday essentials business.

Joining us on the call today for your questions. A replay will be available on our Investor Relations website for at least 3 months. We appreciate your interest in Amazon and look forward to speaking with you again next quarter.

Thank you for standing by. Good day, everyone, and welcome to the Amazon.com First Quarter 2026 Financial Results Teleconference. [Operator Instructions] Today's call is being recorded. And for opening remarks, I will be turning the call over to the Vice President of Investor Relations, Mr. Dave Fildes. Thank you, sir. Please go ahead.

Hello, and welcome to our Q1 2026 financial results conference call. Joining us today to answer your questions is Andy Jassy, our CEO; and Brian Olsavsky, our CFO.

As you listen to today's conference call, we encourage you to have our press release in front of you, which includes our financial results as well as metrics and commentary on the quarter. Please note, unless otherwise stated, all comparisons in this call will be against our results for the comparable period of 2025. Our comments and responses to your questions reflect management's views as of today, April 29, 2026, only, and will include forward-looking statements. Actual results may differ materially. Additional information about factors that could potentially impact our financial results is included in today's press release and our filings with the SEC, including our most recent annual report on Form 10-K and subsequent filings.

During this call, we may discuss certain non-GAAP financial measures. In our press release, slides accompanying this webcast and our filings with the SEC, each of which is posted on our IR website, you will find additional disclosures regarding these non-GAAP measures, including reconciliations of these measures with comparable GAAP measures.

Our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our results are inherently unpredictable and may be materially affected by many factors, including fluctuations in foreign exchange rates and energy prices, changes in global economic and geopolitical conditions, tariff and trade policies, resource and supply volatility, including for memory chips, and customer demand and spending, including the impact of recessionary fears, inflation, interest rates, regional labor market constraints, world events, the rate of growth of the Internet, online commerce, cloud services and new and emerging technologies and the various factors detailed in our filings with the SEC.

Our guidance assumes, among other things, that we don't conclude any additional business acquisitions, restructurings or legal settlements. It's not possible to accurately predict demand for our goods and services, and therefore, our actual results could differ materially from our guidance.

And now I'll turn the call over to Andy.

Thanks, Dave. We're reporting $181.5 billion in revenue, up 17% year-over-year. Excluding the $2.9 billion favorable impact from foreign exchange, net sales increased 15%. Operating income was $23.9 billion.

Q1 was a strong quarter for Amazon. Starting with AWS, growth continued to accelerate, up 28% year-over-year, the fastest growth rate in 15 quarters, up $2 billion quarter-over-quarter, the largest Q4 to Q1 AWS revenue increase ever. AWS is now a $150 billion annualized revenue run rate business. It's very unusual for a business to grow this fast on a base this large. And the last time we saw growth at this clip, AWS was roughly half the size.

We've never seen a technology grow as rapidly as AI. Amazon is already a leader and companies continue to choose AWS for AI. To put our growth in perspective, 3 years after AWS launched, we had a $58 million revenue run rate. In the first 3 years of this AI wave, AWS' AI revenue run rate is over $15 billion, nearly 260x larger.

There are several reasons customers are choosing AWS for AI. First, we've built broader capabilities than others. That includes model building with SageMaker, which reduces training time by up to 40%; high-performance inference with the leading selection of frontier models in Bedrock, which saw 170% growth in customer spend quarter-over-quarter; and processed more tokens in Q1 than all prior years combined.

We're excited to make OpenAI's models available in Bedrock. Yesterday, we added OpenAI's GPT 5.4 model with 5.5 coming soon. Yesterday, we also started the preview of Amazon Bedrock managed agents powered by OpenAI, a stateful run time environment that enables any organization to build generative AI applications and agents at production scale. We believe that modern agentic applications will be stateful, and this new technology will rapidly accelerate Agentic AI adoption. OpenAI has said they're already seeing unprecedented demand for this new product, and we're seeing heavy customer interest as well.

Most of the value companies derive from AI will be through agents, and AWS customers can build agents with their proprietary data and strands, which has been downloaded more than 25 million times and saw 3x more downloads quarter-over-quarter. Customers can deploy agents with enterprise scale, security and reliability with AgentCore, which is being used to deploy an agent as frequently as every 10 seconds. We also offer turnkey agents for coding software migrations, business operations and knowledge workers in Kiro, Transform, Connect and Quick, and they continue to resonate with customers. The number of developers using Kiro more than doubled quarter-over-quarter, and enterprise customer usage increased nearly 10x. Customers have used Transform to save over 1.56 million hours of manual effort when migrating and modernizing their workloads. A number of new customers using Quick has grown more than 4x quarter-over-quarter, and we just announced day 1 of our Quick desktop app yesterday. It's very compelling as it can query your e-mail, calendar, Slack, local files and several other applications you use every day to flag important communications, retrieve and summarize information, make recommendations, compose and send communications to others and create agents that highlight or automatically do work that you used to have to do yourself. You can easily keep refining your preferences, and Quick's advanced Knowledge Graph enables the AI agents to automatically learn from your interactions to become more personalized over time. One of our enterprise customers just told us Quick isn't just improving how we work, it's letting us reimagine it.

Second and another reason customers continue choosing AWS is that as they expand their use of AI, they want their inference to reside near their other applications and data and much more of it resides in AWS than any place else. Third, as customers expand their AI usage, they also want to consume additional non-AI services, and they're choosing AWS because we've built the broadest and most capable core offerings by a wide margin. We offer thousands of features across compute, storage, databases, analytics, security and more, and Gartner consistently recognizes AWS' leadership across their major cloud evaluation areas. Fourth, AWS is the strongest security and operational performance of any AI and infrastructure provider. And start-ups, enterprises and governments continue to choose AWS as the foundation for their most critical workloads.

These are some of the reasons even more customers are choosing AWS. And just since last quarter's call, we've announced new agreements with OpenAI, Anthropic, Meta, NVIDIA, Uber, U.S. Bank, Fox, Southwest Airlines, U.S. Army, Bloomberg, Cerebrus, AT&T, Nokia, Fundamental, The National Geographic Society, PGA Tour and many more.

Our chips business continues to grow rapidly and is larger than what a lot of folks thought. We saw nearly 40% quarter-over-quarter growth in Q1, and our annual revenue run rate is now over $20 billion and growing triple-digit percentages year-over-year. But this somewhat masks the size. If our chips business was a stand-alone business and sold ships produced this year to AWS and other third parties as other leading chip companies do, our annual revenue run rate would be $50 billion. As best as we can tell, our custom silicon business is now one of the top 3 data center chip businesses in the world. The speed at which we've gotten here is extraordinary.

And we have momentum. For our custom AI silicon, we've recently shared very large multiyear multi-gigawatt training commitments from the 2 leading AI labs in the world, Anthropic and OpenAI, as well as an increasing number of companies like Uber betting on Trainium. And we now have over $225 million in revenue commitments for Trainium. Our Trainium 2 chip has about 30% better price performance than comparable GPUs and is largely sold out. Trainium 3, which just started shipping at the start of 2026 and was 30% to 40% more price performance than Trainium, 2 is nearly fully subscribed. And Trainium 4, which is still about 18 months from broad availability, has already been reserved. Amazon Bedrock, which is used expansively by over 125,000 customers, runs most of its inference on Trainium. Almost 80% of the Fortune 100 companies are using Bedrock.

We also just announced that Meta is committed to using tens of millions of Graviton cores. Graviton is our industry-leading CPU chip, which allows Meta to run the CPU intensive workloads behind agentic AI with the performance and efficiency they need at their scale. AI is commonly seen as a GPU story. But the rise of agentic workloads, real-time reasoning, cogeneration, reinforcement learning and multistep task orchestration is driving massive CPU demand as well. As AI systems shift from answering questions to taking actions, it is post training an inference scale up the compute required pulls heavily on CPUs. That's why Meta chose Graviton, which delivers up to 40% better price performance than any other x86 processors, and now used by 98% of the top 1,000 EC2 customers. Nobody has a better set of chips across AI and CPU workloads than AWS with Trainium and Graviton. And we're unusually well positioned for this AI inflection we're in the early stages of experiencing.

While the largest number of AI chips we're bringing in are Trainium, we continue to have a deep partnership with NVIDIA. We have immense respect for them, continue to order substantial quantities. We'll be partners for as long as I can foresee, and we'll always have customers who want to run NVIDIA on AWS. And we will also have a very large chips business ourselves. Customers always want choice. It's always been true and always will be true. Different companies will offer different benefits for customers. And the uniquely strong price performance that Trainium offers is compelling to our external and internal customers.

For perspective, at scale, we expect Trainium will save us tens of billions of dollars of CapEx each year and provide several hundred basis points of operating margin advantage versus relying on other chips for inference.

Finally, we continue to be confident in the long-term CapEx investments we're making. Of the AWS CapEx we intend to spend in 2026 much of which will be installed in future years, we have high confidence that this will be monetized well as we already have customer commitments for a substantial portion of it and that it will yield compelling operating margins and ROIC. As we've been sharing, the faster AWS grows, the more short-term CapEx we'll spend. AWS has to lay out cash for land, power, buildings, chips, servers and networking gear in advance of when we can monetize it, typically 6 to 24 months before we start building customers depending on the component. However, these CapEx investments fund assets with many year useful lives, 30-plus years for data centers, 5 to 6 years for chips, servers and networking gear. The free cash flow and ROIC for these investments are cumulatively quite attractive a couple of years after being in service. However, in times of very high growth like now, where the CapEx growth meaningfully outpaces the revenue growth, the early year's free cash flow is challenged until these initial tranches of capacity are being monetized and revenue growth outpaces CapEx growth. We've been through this cycle with the first big AWS growth wave and like the results. We expect to feel similarly about this next wave with much larger potential downstream revenue and free cash flow.

I'll now turn to stores. Units grew 15% year-over-year, and the highest we've seen since the tail end of COVID lockdowns. We continued expanding selection, including more than 600 new notable brands. Our grocery business continues to grow quickly across both perishables and nonperishables. And with more than $150 billion in gross sales in 2025, we're now the second largest grocer in the U.S. We offer perishables delivered same day alongside millions of other items in more than 2,300 cities and towns across the U.S. with more to come. Prime members are loving the convenience of getting fresh groceries alongside other products they're buying in Amazon. And perishable sales have grown over 40x year-over-year, and make up 9 of the top 10 most ordered items for same-day delivery where the service is available. Customers shopping same-day perishables build larger baskets, adding nearly 3x as many items to their order, and has spent over 80% more than customers who don't.

Whole Foods market also continues to accelerate with over 550 stores today and 100 more coming in the next few years. We remain committed to meeting or beating other retailers on price. And in Q1, the average prices of products offered on Amazon.com decreased compared to the same period last year. Prime Day will take place in most countries in June, which will bring Prime members even more savings across every category.

We continue to find new ways to speed up delivery for customers in both cities and rural areas. We offer millions of items available for same-day delivery with Prime, up to 40x the selection of a typical big box retail store, and we've delivered more than 1 billion items same-day overnight so far this year.

We're also making delivery even faster, recently announcing 1- and 3-hour delivery options on over 90,000 items with 1-hour delivery available in hundreds of cities and towns, 3-hour delivery in 2,000-plus cities and towns and more on the way. And we continue to expand our ultrafast delivery service, Amazon Now, which offers delivery in 30 minutes or less on thousands of items. It started last year in India, where orders are increasing 25% month-over-month with Prime members tripling their shopping frequency once they start using it. The service is now available to tens of millions of customers across 9 countries with more to come as well.

The stores team also continues to innovate and deliver for customers with AI. We launched Health AI, a 24/7 AI-powered personal health agent backed by One Medical clinicians that gives U.S. customers instant clinical guidance and takes action with their permission, from booking appointments to managing prescriptions to facilitating medical treatment with a real One Medical provider. Rufus, our agentic AI shopping assistant, continues to resonate with customers. Rufus can research products, track prices and auto buy products in our store when they reach a set price. Monthly active users are up over 115%, and engagement is up nearly 400% year-over-year.

And we recently introduced a new AI experience for sellers at Seller Central that dynamically generates a custom personalized visualization of data, key insights and scenarios tailored to the seller's goals. It's early, but the initial response and feedback are very strong.

Moving on to Amazon Ads. We continue working to be the best place for brands of all sizes to grow their businesses, and we're pleased with the continued strong growth across our full funnel offerings, generating $17.2 billion of revenue in the quarter and up 22% year-over-year. Forrester recently recognized Amazon Ads as a leader in omnichannel advertising platforms with unmatched supply and insights for connected TV and commerce media. We deepened our Netflix partnership with Amazon Audiences, which enables advertisers to apply Amazon's exclusive signals from shopping, browsing and streaming to Netflix's highly engaged viewers to reach the right audiences and drive even stronger performance. We also partnered with Comcast Advertising to expand local advertising to thousands of brands and expanded interactive video ad capabilities to partners starting with Samsung TVs.

Our ads team also continues to invent and deliver for advertisers with AI. For example, we expanded creative agent, an agentic partner that plans and executes the entire ad creative process to Canada, France, Germany, India, Italy, Spain and the U.K. And we recently introduced sponsored product and brand prompts in Rufus to help brands showcase products and customers make more informed buying decisions. It's early but we're seeing nearly 20% of shoppers who interact with the brand prompt in Rufus continue the conversation about that brand.

We're also continuing to invent and see momentum in several other areas. I'll mention a few. Starting with entertainment. Moviegoers have flocked to Project Hail Mary with nearly $615 million in global box office to date. Its opening weekend was the second biggest for any non-sequel non-franchise film in the last decade. We also surpassed 100 million viewers globally for the Copolla's movie Trilogy, with all 3 films reaching #1 in more than 170 countries at launch. In live sports, we offered exclusive coverage of the NBA SoFi play with total viewership up 18% compared to last year on cable.

Alexa Plus early access expanded to millions more Prime members in Mexico, the U.K., Italy and Spain. Customers are loving Alexa Plus, talking to Alexa twice as much and for longer durations across a wider breadth of topics, completing purchases on devices 3x more, streaming music 25% more and using smart home functionality 50% more than Alexa Classic.

Zoox has now driven nearly 2 million miles and carried more than 350,000 riders. It's available to the public in Las Vegas and San Francisco, and is testing in 8 other cities. We recently announced that Zoox will be available through the Uber app in Las Vegas and Los Angeles in the future.

And finally, Amazon Leo continues gaining momentum with commercial service on track to launch in a few months. We already have meaningful revenue commitments from enterprises and governments, including Delta Airlines, JetBlue, AT&T, Vodafone, DIRECTV Latin America, Australia's national broadband network, DP World Tour, NASA and others. We also announced that we plan to acquire Globalstar, which will expand Leo's satellite network with direct-to-device capabilities. And we entered an agreement with Apple for Amazon Leo to power satellite services for iPhones and Apple Watches.

We're in the middle of some of the biggest inflections of our lifetime. And Amazon has the culture, the know-how and the resources to make so many customers' lives better and easier and to build multiple new long-term businesses with substantial return on invested capital and free cash flow. We will continue investing in inventing to make it so.

With that, I'll turn it over to Brian.

Thanks, Andy. Let's start with our top line financial results. Worldwide revenue was $181.5 billion, a 15% increase year-over-year, excluding the 180 basis point favorable impact of foreign exchange. Worldwide operating income was $23.9 billion with an operating margin of 13.1%, our highest operating margin ever.

Across all segments, we continue to innovate for customers while operating more efficiently. In the North America segment, first quarter revenue was $104.1 billion, an increase of 12% year-over-year. International segment revenue was $39.8 billion, an increase of 11% year-over-year, excluding the impact of foreign exchange. The seasonal shopping events performed well in Q1, including our big spring sale. We also saw particularly strong performance for third-party sellers for important contributors to our broad selection and competitive pricing. Our sellers saw strong sales growth in Q1, particularly in the U.S. as well as in Europe and Brazil, where we've recently lowered seller fees. We're seeing our investments in the seller experience resonate and in turn, grow our business.

Prime continues to fuel our growth and reflects the value members received from the program. Prime Video is a key pillar of the Prime value proposition and an important driver of new member acquisition. Our investments in Original and Exclusive Content and Live Sports, combined with our third-party partner titles, offer the best selection of premium video content. In addition to delivering compelling value to Prime members, advertisers and partners, Prime Video is now large and profitable business in its own right.

Now let's shift to segment profitability. North America segment operating income was $8.3 billion with an operating margin of 7.9%. International segment operating income was $1.4 billion, with an operating margin of 3.6%. We are pleased with the fulfillment network performance in Q1. The team has worked hard to optimize our network. Overall unit growth of 15% continues to outpace our cost to operate the fulfillment network as outbound shipping costs grew 12% year-over-year and fulfillment expense grew 9% year-over-year, both on an FX-neutral basis. As our network efficiency improves, we're able to deliver items faster and improve the customer experience, while at the same time lowering our cost to serve.

Looking ahead, we see meaningful opportunities to further enhance productivity across our global fulfillment network, all while continuing to raise the bar and delivery speed. We will keep optimizing inventory placement to shorten distance traveled, reduce touches per package and improve consolidation rates.

Alongside these efforts, we deploy robotics and automation, which have been integral to our operations for decades. Our latest generation technologies offer a step change in efficiency, which we're deploying in both new and existing facilities. All of our U.S. large-format fulfillment center launches in 2026 will have this latest generation technology. We're seeing early positive results with improved site safety, higher productivity and lower cost to serve.

Moving to our AWS segment. Revenue was $37.6 billion, and growth accelerated 480 basis points to 28% year-over-year, driven by both core and AI services. We continue to see customers increase cloud migrations and scale their use of AWS core services. Customers seeking the full benefit of AI are accelerating their transition to the cloud. We also see a strong correlation between AI spend and core growth. If customers spend more on AI, we see a corresponding demand increase in core. We expect this to increase over time as customers move more AI workloads into production, strengthening demand for our core services. Our AI revenue is growing triple digits year-over-year. We're bringing more capacity online to meet high customer demand while also driving meaningful efficiency gains across our installed base.

Our AI offerings continue to gain traction with customers, and Bedrock has been a significant growth driver. In 2025, we delivered 4x improvements in Trainium 2's token throughput. Since the majority of Bedrock's workloads run in Trainium, these efficiency gains directly translate into more capacity to serve customers. AWS operating income was $14.2 billion and reflects our strong growth, coupled with our focus on driving efficiencies across the business.

Now turning to total company capital expenditures. Our cash CapEx is $43.2 billion in Q1. This primarily relates to AWS and generative AI as we invest to support strong customer demand. We'll continue to make significant investments, especially in AI, as we believe it to be a massive opportunity with the potential to drive long-term revenue and free cash flow.

I'll finish with our financial guidance for Q2. The following guidance assumes that Prime Day occurs in the second quarter in most of our largest geographies, including the U.S., and that Prime Day occurs in the third quarter in Australia, Brazil, India and Japan. Note that in 2025, Prime Day was in Q3 for all countries.

Q2 net sales are expected to be between $194 billion and $199 billion. We estimate the year-over-year impact of changes in foreign exchange rates based on current rates, which we expect to be a headwind of approximately 10 basis points in the quarter. Q2 operating income is expected to be between $20 billion and $24 billion. We continue to see strong sales trends carrying into Q2, and I'll mention a few items on the operating income guidance.

First, this estimate includes the impact of our seasonal step-up in stock-based compensation expense in Q2 driven by the timing of our annual compensation cycle.

Second, within the North America segment, we do expect a year-over-year cost increase of approximately $1 billion related to Amazon Leo as we manufacture and launch more satellites in preparation for our service offering. Amazon Leo's commercial service is on track to launch in Q3, and we expect to begin capitalizing certain costs in Q4, including production and launch costs.

Third, our guidance anticipates higher transportation costs related to fuel inflation, which is partially offset by the recently implemented fuel and logistics-related FBA surcharge.

I'm thankful to our teams across the company for their hard work and dedication to customers. We remain focused on driving an even better customer experience, which is the only reliable way to create lasting value for our shareholders. With that, let's move on to your questions.

Thank you. At this time, we will now open the call up for questions. [Operator Instructions] And the first question comes from the line of Eric Sheridan with Goldman Sachs.

Andy, across an array of announcements you've made recently with AWS and reflecting upon what you wrote in the shareholder letter, can you talk a little bit about the needed levels of investment over the next couple of years to scale compute and capacity to meet your current state of revenue backlog and how we should be thinking about your unique approach to custom silicon and AI infrastructure that maybe positions you competitively to build that scale?

Yes. Well, to your point, Eric, we've made a lot of announcements over the last several months, and we're really pleased with the growth that we're seeing in AWS right now, 28% year-over-year, fastest growth rate in 15 quarters for us, haven't grown at this pace since we're about half the size. And growing 28% on a $150 billion annual run rate basis is not simple to do. And I think there's a few things around it.

First is just we continue to see people choosing AWS for AI in part because of our really broad full stack functionality, in part because people want their inferences, they scale it to be close to their data and their applications so much more that it lives in AWS and elsewhere. And in part because we have a strong security and operational performance. And that's just -- you can see it in our numbers, it's leading to very substantial AI growth. And then at the same time, we're seeing very significant growth in our core business. And some of that are the migrations that have picked up from enterprises from on-premises to the cloud. But a lot of that is also as AI growth is exploding, it turns out that it leads to a lot of core growth as well. All the post training, all the reinforcement learning, all the agentic actions and tool usage that these agents are using, and it fits with what you're asking about on the chip side, which is because we have an unusual collection of chips, we have the leading CPU chip and Graviton and we have the leading price performance silicon, AI chip and Trainium, it means that we're really unusually well positioned for the inflection that we're seeing and the type of growth that we're experiencing.

And so I don't have an update on -- a new update on capital. Our plan is largely the same, but we do view this as truly a once-in-a-lifetime opportunity where every application that we know of is going to be reinvented. And there are so many new applications that none of us have ever imagined or dreamed we could build, that are starting to be built and will be built. And all of that is going to be built on top of AI with a lot of consumption of CPUs and core as well. So I think -- I expect that we will invest a significant amount of capital over the coming years to pursue that opportunity. And that our customers, our shareholders and Amazon in general are going to be much better off down the road because we did so.

And the next question comes from the line of Brian Nowak with Morgan Stanley.

I have 2. One is on the accounting side, we'll probably get it in the queue, but can you just give us an update on what the AWS backlog looks like? And sort of any visibility on the breadth of that backlog beyond the big labs? That's the first one.

And then the second one. As you sort of think about milestones for Rufus and agentic commerce for you in 2026, what are you most focused on making sure you accomplish on the agentic side this year just to make sure you stay at the nice edge of the agentic commerce offerings?

Yes. On the backlog, the backlog for Q1 is $364 billion. That does not include the recent deal that we announced with Anthropic for over $100 billion. There's reasonable breadth in that as well. It's not just one customer or 2 customers.

On the agentic commerce milestone question. We are very bullish on what agentic commerce will look like. I think it's going to be very good for customers in the long term. I think it will be good for us, too. And you can see some of that focus from us and what we're building with Rufus. If you haven't checked out Rufus in a while, it's really substantially improved over the last year, and we have a lot of customers using it. As I mentioned, earlier, you see the monthly active users up over 115% in Rufus and the engagement up over 400% year-over-year. And I think while I think there will be -- we'll do a lot of work with third-party horizontal agents to try and make that customer experience better -- and by the way, I do think today, it reminds me in some ways the stage we're in of what we saw in the early days of search engines and they're trying to refer business to e-commerce. It's never been a giant part of the referrals to our e-commerce business. But over the years, the experience got better. And what you see with agentic commerce is it's a small fraction of what we see with the search engine referrals, but the experience just hasn't gotten great with these third-party horizontal agents yet. They're not often able to get the pricing right or the product information right. They don't have any personalization data or any shopping history. And so we do want to see that get better with third-party horizontal agents. We're having conversations with all those folks to try and make that better and find something that works for customers and all the companies. And then it will be interesting over time which agents customers choose to use.

I happen to think that if you're going to a particular retailer that you'd like to do business with and you like to shop from, if they have a great agentic shopping assistant, you're going to often start there because it's where you're doing your shopping, it's easier to -- they have better product information, they have better information about what are the customers like you are buying. You can make all sorts of changes to how your account and your shipping information is working there. And so that's what we're aiming to make Rufus be is we're aiming to have it be the best shopping assistant anywhere, and I think we're on that path.

The next question comes from the line of Justin Post with Bank of America.

I'd like to ask 2. One on models and then one on Trainium chips. So on models, it looks like you might have access to the full suite of OpenAI models on Bedrock. Just wondering how big of an unlock that is and how focused maybe you are on your own Nova model?

And then second, shareholder letter mentioned you might be able to sell racks of Trainium. Just wondering, with your capacity constraints, how you think about timing of that? And how big of an opportunity?

Yes. On the models question, I think the fact that we're going to have all the OpenAI models available in Bedrock is a big deal. It's a big deal for customers. And we have -- we obviously have a very large amount of AI being done in Bedrock today on the models we have. And this is Anthropic and [ Lama and Strong ] and a host of others. But the one thing you learn over and over again with every technology, it was true in databases, it was true in analytics, it's true in models. It's true in chips too, by the way, is that customers want choice. There is not one tool to rule the world, and they want choice. And each of the models are better at some things than the other models. And so people for a long time have wanted to consume OpenAI models in Bedrock. We just enabled yesterday the stateless model, the 5.4 model and will enable the most recent 5.5 model in the next couple of weeks. And most of the model work and most of the AI has been done in these stateless models, kind of tokens in and tokens out. And while I think there will continue to be a lot of work on that way, I think the future of using these models is a stateful model, a stateful API. And that's because when you're building agents, you're building AI applications, you don't want to start anew every time you interact with the model. You want to store state, you want to you want to store identity, you want to store what the conversation or the actions have been. You want to reach out and do a little bit of compute here. You want to have the tools to be able to reach -- the models reach out to the different tools to accomplish different tasks. And that only happens if you're able to store state. And so the Bedrock-managed agents that we collaborate with and invented with OpenAI that we just announced a preview of yesterday is also -- I think that's the future of how these agents are going to be built. It's something that nobody else has. And I think it's very exciting to our customers. And of course, we'll have other models like codecs and things like that as well. So I think it's a big deal for customers, and I think it's going to be good for our business as well.

On the question about Trainium and the notion of our selling racks over time, I do think that's very much a possibility. Always, we have to balance -- we have such demand right now for Trainium, and we have such demand from various companies who will consume as much as we make that we have to decide how much we're going to allocate to the existing demand and customers how much we're going to save to sell these racks. And for our existing customers that we sell Trainium to, how many would be Trainium plus running on our cloud infrastructure versus just the chips themselves. But I expect over time, there's a good chance we're going to sell racks over the next couple of years.

And the next question comes from the line of Rob Sanderson with Loop Capital Markets.

I wanted to ask a little bit about Amazon Leo. Can you maybe help dimensionalize some of the revenue opportunity in the consumer and in the enterprise space over the next few years. What are the governors on the ramp? Could you talk about types of new services that you will be able to develop with the Global Star infrastructure and the spectrum that maybe you couldn't address before or would take you, you can get to more quickly now.

And then how expansive is the longer-term vision? I know you're just beginning to launch commercial services, but over the long term, do you expect to include noncommunication services like -- data centers or things like that as this becomes feasible in the decade ahead?

Yes, I'll try and address as many of those questions as I can. I am very bullish about Amazon Leo and the opportunity there. There are billions of people around the world who do not have access to broadband connectivity. And there are many thousands of businesses and government assets that just -- that people don't have visibility to because they don't have the right connectivity. And it means that those entities can't do a lot of the things that we all take for granted today, including education online business online, shopping or entertainment online, having constant visibility and digital twins. There's all these things that they can't do today. And so we think that Amazon Leo is going to help solve that problem. I think when we launched our service commercially, we've got -- we just had another launch this week, so we have over 250 satellites in space. When we launched that service commercially, it will be one of 2 offerings that are on the current technology edge. And I think that we will have a meaningful advantage and performance, I think, will be about 2x better on the downlink than existing alternatives at about 6x better on the uplink performance than existing alternatives. I think we'll have a cost advantage for customers.

And then for the governments and the enterprises, and we talk to a lot of them, and we have already signed agreements with many of them, even though we haven't launched the service commercially, the latest of which was Delta Airlines committing at least half of their fleet starting in 2028. When you talk to them another really big part of what matters them is they're going to want to take this data off of the satellite constellation and they're going to want to store it in the cloud and they're going to want to do analytics on it and they're going to want to do AI on it. And just the combination of Leo with the leading cloud in the world in AWS is very compelling to enterprises and to government.

So I think the are only -- today, if you ask what stops us from growing the business, we have to get the constellation into space. We have over 20 launches planned. This year, we have over 30 launches planned in 2027. But I think the business has a chance to be a very large, many billion dollar revenue business. And I think it has some characteristics that are reminiscent of AWS in that it's capital intensive upfront, where you're committing a lot of capital and cash in the early years for assets that you get to leverage over a long period of time. And so I like the free cash flow and return on invested capital characteristics of that business in the medium to long term.

And the last thing I'll say about it is your question about Globalstar. Increasingly, what we're finding with consumers and enterprise and governments is that they don't like to have any periods where they don't have connectivity. It just upsets whatever customer experience they're going through. Even in metropolitan areas, we all hit certain parts of the highway or certain roads where you can't get connectivity or you're hiking, you're skiing. And so increasingly, we see very large demand for consumers to have direct to device. And that was really the impetus for our acquisition of Globalstar. They have unusual and scarce global spectrum that's required to provide direct to device. We also really like the satellite know-how that we'll get as part of that merger with Globalstar. And then it also afforded us the opportunity to build a deep relationship with Apple, who is going to use our direct to device for their iPhones and for their Watches. So very optimistic about the business.

And the next question comes from the line of Shweta Khajuria with Wolfe Research.

I wonder, Andy, if you could please talk about how you're thinking about the increase in price for memory and storage and just the supply chain inflation we're seeing and the impact it could have in CapEx this year and potentially next year as well?

And then on agentic commerce, if you could talk about how you view the opportunity with advertising. I have no doubt that Rufus could be the best shopping assistant available over time. But for advertising opportunity, how do you view that if agents would be the ones taking action to shop

So on memory and storage and the supply chain, I think everybody knows that the cost of these components, particularly of memory, have skyrocketed, and we're just in a stage where there's just not enough capacity for the amount of demand. We have worked very closely with our strategic partners. We saw this trend happening early in the kind of the middle of the latter part of last year, and we've worked with our strategic suppliers here to get a significant amount of supply. And so we're working very closely with them. I think the team has been very scrappy. I think we've done a good job in making sure that we're not capacity constrained there, but we watch that very closely.

One of the interesting things that we see right now with the change in price and in supply on things like memory is that it is a further impetus pushing companies who have on-premises infrastructure into the cloud. And it's because of a meaningful part, the suppliers are prioritizing their very largest customers, which cloud providers are. And so we have seen a number of conversations we've been having with enterprises for many months where it's just been slower in getting the transformation plan to move to the cloud, accelerate rapidly just because we have a lot more supply than what others have. So it'd be interesting to see how that evolves over time. It could have -- we're doing our best to kind of -- to have the supply we need and keep the cost in the right spot, but we'll see how that continues to evolve.

And I think on the agentic commerce and how that impacts advertising, I actually believe that we're going to we're going to like this for advertising. I think it's going to be good for customers, and it's going to be good for our business. And I think, first of all, the first thing to remember is the way that our ads team has built tools and agents themselves is making it so much easier to do advertising. If you look at small and medium-sized businesses that had to take weeks and months to do creative and to pick the right audience, but all of that is just -- it's so much faster and so much easier because of our advertising agentic tools. And you no longer have to take as much time or spend as much money building the creative. So I think they're going to be a lot more advertising -- advertisers with the rise of what's happening in AI.

And then if you look at the agentic commerce experiences, and you look at any of these agentic experiences, they tend to be multiturn conversations where you're not interacting with one search and getting an answer, you tend to find that you're asking questions, you're narrowing questions, it's asking you questions on what you want. And in that process of having multi turns, there are multiple opportunities to surface relevant products to customers, many of which will be organic and some of which will be sponsored. And it also gives rise to opportunities like sponsored prompts. And so one of the interesting things that has been very successful for customers in our store has been when they ask certain questions, we give them a number of suggestions that are all created through AI. And we've gotten pretty good also having sponsored prompts and that mix of questions and prompts to make it easy for people to keep digging deeper into what they're interested in. So I actually believe that advertising will do well in a world of agentic commerce.

Thank you. And our final question comes from the line of Colin Sebastian with Baird.

Maybe a 2-parter, if I could. Andy, first off, just wondering where you're seeing in terms of the trend between incremental AI demand from earlier adopters and larger AWS customers versus maybe how the demand curve is shaping up across the broader enterprise base?

And then at a high level, as you think about the use of AI internally across Amazon's businesses, presumably the business overall looks very different in 3 or 4 years. Maybe, Andy, if you could contextualize where you see the most opportunity for the technology internally, both in terms of product as well as maybe driving more operating efficiency, I think that would be helpful.

Yes. So on the -- what we see in the incremental AI demand from early adopters versus broader enterprise base, there's -- I think it's no secret that you've got the AI labs are spending an incredible amount of money on compute at this point -- in compute, both on the AI side as well as on the core side. And the models that they're building and the companies that have successful generative AI applications are certainly spending a lot. And there's several of those labs, but we also see quite a bit of enterprise adoption and usage of AI. As I've said before, the gorgeous absolute place that we see enterprises having success is in projects that are around cost avoidance and productivity. These are things like automating customer service or business process automation or fraud or things of that sort. But the number of projects that we're working with across enterprises and that we're now starting to see come to production around brand-new experiences, trying to figure out how to reinvent their current experiences but using inference and AI to be smarter -- also very significant. So we're seeing the adoption of both of those segments.

On the use of AI internally and for our current businesses, I think that the shortest first summary I could give you, Colin, is that I do not see a place in any of our businesses or any of the ways that we do work where we're not going to have giant impact on what we do. I think I've long had this belief that while you can add incrementally to a lot of your existing customer experiences, different agentic and AI experiences. I really believe that in the fullness of time, and I don't know if that's 3 years from now or 5 years from now or it could be sooner, too, that all these customer experiences we know are going to be completely reinvented. And they're going to have different interfaces. They're going to have different ways that people interact with them. They're going to -- people are going to want to have dialogue with them. And so I think it means that you have to look -- it's tricky for if you have an existing business that's doing well. But you have to look at every single one of your customer experiences and you have to be able to carve off resource for that team to think anew about what would the future customer experience look like. If you start from scratch today and if you had all the technologies like AI available to you when you start it. And that is what we're doing in every single one of our experiences. And if I have a chance to be involved in some of those, and it's really exciting. And there are experiences that may take a while to -- for customers to get used to and to use over time. You might find different segments like those AI-forward experiences more than others early on. But if you're not actually working on inventing those right now, I think it's going to be very hard to have the business and the experience leadership that we want over a long period of time. So every single one of our consumer businesses, every single one of our businesses in general is working on that.

And then I would say internally, I also think that it's going to radically change how we work. It already is. I mean just look at how coding, agentic coding is changing how we're all building products. I think it's going to have a comparable impact on how we do DevOps and how we do customer service, how we do research, how we do analytics, how sales is conducted. I think every single one of these functions that we all do at work are going to very significantly change. And that's another area of real focus for us.

And we have this experience. I mentioned in my letter, but if you look at one of our services, we swapped out the engine of the service while we are also running the service full tilt. And normally, that would have taken 40 or 50 people about a year to do, and we took 5 really smart people, AI forward-thinking people building on agentic coding tools, and those 5 people rebuilt it in 65 days. Like that is a very different world of operating, and that's the world I think we're heading to over the next few years.

Thanks for joining us on the call today and for your questions. A replay will be available on our Investor Relations website for at least 3 months. We appreciate your interest in Amazon and look forward to talking with you again next quarter.

Thank you for standing by. Good day, everyone, and welcome to the Amazon.com Fourth Quarter 2025 Financial Results Teleconference. [Operator Instructions] Today's call is being recorded.

And for opening remarks, I will be turning the call over to the Vice President of Investor Relations, Mr. Dave Fildes. Thank you, sir. Please go ahead.

Hello, and welcome to our Q4 2025 financial results conference call. Joining us today to answer your questions is Andy Jassy, our CEO; and Brian Olsavsky, our CFO.

As you listen to today's conference call, we encourage you to have our press release in front of you, which includes our financial results as well as metrics and commentary on the quarter. Please note, unless otherwise stated, all comparisons in this call will be against our results for the comparable period of 2024. Our comments and responses to your questions reflect management's views as of today, February 5, 2026, only, and will include forward-looking statements. Actual results may differ materially. Additional information about factors that could potentially impact our financial results is included in today's press release and our filings with the SEC, including our most recent annual report on Form 10-K and subsequent filings.

During this call, we may discuss certain non-GAAP financial measures. In our press release, slides accompanying this webcast and our filings with the SEC, each of which is posted on our IR website, you will find additional disclosures regarding these non-GAAP measures, including reconciliations of these measures with comparable GAAP measures.

Our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our results are inherently unpredictable and may be materially affected by many factors, including fluctuations in foreign exchange rates and energy prices, changes in global economic and geopolitical conditions, tariff and trade policies, resource and supply volatility, including for memory chips and customer demand and spending, including the impact of recessionary fears, inflation, interest rates, regional labor market constraints, world events, the rate of growth of the Internet, online commerce, cloud services and new and emerging technologies and the various factors detailed in our filings with the SEC. Our guidance assumes, among other things, that we don't conclude any additional business acquisitions, restructurings or legal settlements. It's not possible to accurately predict demand for our goods and services, and therefore, our actual results could differ materially from our guidance.

And now I'll turn the call over to Andy.

Thanks, Dave. We're reporting $213.4 billion in revenue, up 12% year-over-year, excluding the impact from foreign exchange rates. Operating income was $25 billion and trailing 12-month free cash flow was $11.2 billion.

We're seeing strong growth and with the incremental opportunities available to us in areas like AI, chips, low earth orbit satellites, quick commerce and serving more consumers' everyday essentials needs, we have a chance to build an even more meaningful business in Amazon in the coming years with strong return on invested capital, and we're investing to do so.

We're already seeing strong demand in these areas even in these early innings. I'll start with AWS. AWS growth continued to accelerate to 24%, the fastest we've seen in 13 quarters, up $2.6 billion quarter-over-quarter and nearly $7 billion year-over-year. AWS is now a $142 billion annualized run rate business, and our chips business, inclusive of Graviton and Trainium is now over $10 billion in annual revenue run rate, growing triple-digit percentages year-over-year. As a reminder, it's very different having 24% year-over-year growth on a $142 billion annualized run rate than to have a higher percentage growth on a meaningfully smaller base, which is the case with our competitors. We continue to add more incremental revenue and capacity than others and extend our leadership position.

We're continuing to see strong growth in core non-AI workloads as enterprises return to focusing on moving infrastructure from on-premises to the cloud, along with AWS having the broadest functionality, strongest security and operational performance and most vibrant partner ecosystem. AWS continues to earn most of the big enterprise and government transitions to cloud.

Since our last call, we announced new agreements with OpenAI, Visa, the NBA, BlackRock, Perplexity, Lyft, United Airlines, DoorDash, Salesforce, U.S. Air Force, Adobe, Thomson Reuters, AT&T, S&P Global, National Bank of Canada, the London Stock Exchange Group, Choice Hotels, Accenture, Indeed, HSBC, CrowdStrike, and more. More of the top 500 U.S. start-ups use AWS as their primary cloud provider than the next two providers combined.

We're adding significant EC2 core computing capacity each day, and the majority of that new compute is using our custom CPU silicon, Graviton. Graviton is up to 40% more price performance than leading x86 processors and is used expansively by over 90% of AWS' top 1,000 customers. Graviton itself is a multibillion-dollar annualized run rate business, growing more than 50% year-over-year. We consistently see customers wanting to run their AI workloads where the rest of their applications and data are.

We're also seeing that as customers run large AI workloads on AWS, they're adding to their core AWS footprint as well. But the biggest reason that AWS continues to gain AI share is our uniquely broad top to bottom AI stack functionality. In AI, we're doing what we've always done in AWS, solving customer challenges.

Let me give you some examples. The first challenge is having a strong foundation model to generate inferences or predictions. Customers are realizing as they get further into AI that they need choice as different models are better on different dimensions. In fact, most sophisticated AI applications leverage multiple models, whether customers want frontier models like Anthropic's Claude or open models like Mistral or Llama, Frontier Intelligence with lower cost and latency like Amazon Nova or video and audio models like TwelveLabs or Nova Sonic. Amazon Bedrock makes it easy to use these models to run inference securely, scalably and performantly. Bedrock is now a multibillion-dollar annualized run rate business and customer spend grew 60% quarter-over-quarter.

The second challenge is how to hone the model for your application. Customers sometimes think if they have a good model, they will have a good AI application. It's not really true. It takes a lot of work to post train and fine-tune a model for your application. Our SageMaker AI service, along with fine-tuning tools in Bedrock make this much easier for customers.

A third challenge is how to have a custom version of a foundation model that best leverages the company's secret sauce, their own data. To date, companies have tried to shape models with their own data late in the process, usually with fine-tuning or post-training. There's a debate in the industry about this, but we believe that enterprises will want models trained on their own data at an early stage of pretraining if possible. So their models have the best possible foundation for what matters most to each enterprise on which to learn and evolve. It's a little like teaching a child of foreign language early in their life. That becomes part of their learning foundation moving forward, and it makes it easier to pick up other languages later in their life.

To solve for this need, we just launched Nova Forge, which give customers early checkpoints on our Amazon Nova models, allows them to securely mix their own proprietary data with the models data in the pretraining stage and enables their own uniquely customized versions of Nova, what we call Novellas, trained with their data early in the process. This will be very useful for companies as they build their own agents on top of the model. There is nothing else out there like this today and a potential game changer for companies.

Another challenge is cost. I've said this many times, but if we want AI to be used as expansively as companies want, we have to make the cost of inference lower. A significant impediment today is the cost of AI chips. Customers are starving for better price performance. And typically, and understandably, the dominant early leaders aren't in a hurry to make that happen. They have other priorities. It's why we've built our own custom silicon and Trainium, and it's really taken off. We've landed over 1.4 million Trainium2 chips, our fastest ramping chip launch ever. Trainium2 is 30% to 40% more price performance than comparable GPUs and is a multibillion-dollar annualized revenue run rate business with 100,000-plus companies using it as Trainium is the majority underpinning of Bedrock usage today. We recently launched Trainium2, which is up to 40% more price performance than Trainium2. We're seeing very strong demand for Trainium3 and expect nearly all of our Trainium3 supply of chips to be committed by mid-2026. And though we're still building Trainium4, we're seeing very strong interest already.

Looking ahead, the primary way companies will get value from AI is with agents, some their own, some from others, and there are several customer challenges that we're well positioned to solve. It's harder to build agents than it should be. For that, we've built Strands, a service enabling agents to be created from any model. Once agents are built, enterprises are apprehensive about deploying to production because these agents need to securely and scalably connect to compute, data, tools, memory, identity, policy governance, performance monitoring and other elements. This is a new and hard problem where a solution has not existed until we launched Bedrock AgentCore. Customers are quite excited about AgentCore, and it's unlocking deployments.

Customers also want to leverage others' useful agents, and we've built several, including Kiro for coding, Amazon Quick for knowledge workers to leverage their own data and analytics, AWS Transform for software migration and Amazon Connect for call center operations. We continue adding new capabilities and usage continues to grow quickly. For example, the number of developers using Kiro grew more than 150% quarter-over-quarter.

In addition to agents that customers direct, customers are also becoming excited about agents that require less human interaction. They can be fully autonomous, run persistently for hours or days, scale out quickly and remember context. At this past AWS re:Invent, we launched Frontier Agents to do that. Kiro autonomous agents for coding tasks, AWS DevOps agents for detecting and resolving operational issues and AWS security agents for proactively securing applications throughout the development life cycle, and they're already making a big difference for customers.

We expect to invest about $200 billion in capital expenditures across Amazon, but predominantly in AWS because we have very high demand, customers really want AWS for core and AI workloads, and we're monetizing capacity as fast as we can install it. We have deep experience understanding demand signals in the AWS business and then turning that capacity into strong return on invested capital. We're confident this will be the case here as well.

I'll now turn to stores. We continue to expand selection, including more than 400 new beauty brands in the U.S. in 2025, like Bobbi Brown Cosmetics, Charlotte Tilbury and Laura Mercier and new fashion brands like Away Luggage, Converse, Diesel, Michael Kors, Nike and The North Face. Our ultra-low-priced offering, Amazon Haul, grew selection to over 1 million items under $10 and expanded to serve customers in more than 25 countries and regions.

We continue to see strong customer response to Everyday Essentials and grocery. In 2025, Everyday Essentials grew nearly twice as fast as all other categories in the U.S., representing one out of every three units sold in our store, and we've become a go-to grocery destination for over 150 million Americans, mostly through online shopping and Whole Foods. With over $150 billion in gross sales, Amazon is clearly a large grocer at this point. Customers in thousands of U.S. cities and towns can now get perishables delivered same day alongside millions of other items and customers who use that service shop more than twice as often as customers who don't. We plan to expand to many more communities in 2026, and we also plan to open more than 100 new Whole Foods Market stores over the next few years as we work to make grocery shopping easier, faster and more affordable for customers.

We remain committed to staying sharp on price and continue to meet or beat other retailers' prices. A recent study from Profitero showed that Amazon is America's lowest priced retailer for the ninth straight year, 14% lower on average than other major online retailers. We again achieved our fastest ever delivery speeds for Prime members around the world in 2025. In the U.S., we delivered nearly 70% more items same day than the year before. We also continue increasing speed for rural customers with nearly 2x more average monthly customers in rural areas receiving same-day delivery year-over-year.

Same-day is our fastest-growing delivery offering and nearly 100 million customers used it last year in the U.S. And the team is continuing to innovate. We've launched Amazon now in India, Mexico and the UAE, ultrafast delivery on thousands of items in about 30 minutes or less, and we're testing it in several communities in the U.S. and U.K. It's early, but customers are loving it. For example, in India, customer response exceeded our most optimistic expectations, and we're seeing Prime members triple their shopping frequency once they start using it.

Expanding our same-day delivery coverage also leads to meaningfully later cutoff times for orders, which is a big deal for customers. For example, on Christmas Eve, customers in about 4,000 U.S. cities could order items up until midday and get them that same day. Another example is our recently launched feature add to Delivery, which enables Prime members in the U.S. to add items to their upcoming Amazon deliveries with just one tap without going through checkout again or paying additional shipping fees.

Just six months after launch, ad to delivery already makes up about 10% of all Prime volume fulfilled through the Amazon network each week. While this seems simple on the surface, this feature is supported by a lot of invention where we need to figure out in real time and with incredibly low latency, what items among Amazon's hundreds of millions of products are available to add to a customer's upcoming deliveries, surface them, find a way to include in their packages and deliver within the same customer promise.

The stores team is also continuing to innovate and deliver for customers with AI. Our Agentic AI shopping assistant, Rufus, has rapidly expanded. Rufus can research products, track prices and auto buy, purchasing a product in our store when it reaches your set price. It can also now shop tens of millions of items in other online stores and make purchases for customers using our Agentic Buy for Me feature. Last year, more than 300 million customers used Rufus. In addition, customers use Lens, our AI-powered visual search tool to find products with a phone's camera, a screenshot or a barcode, and they did it 45% more year-over-year.

Moving on to Amazon Ads. We're pleased with the continued strong growth across our full funnel offerings, generating $21.3 billion of revenue in the quarter and growing 22% year-over-year. Sponsored products advertising in our store continues to be our largest ads offering and the combination of trillions of shopping, browsing and streaming signals with advanced AI and machine learning led us to deliver highly relevant and useful ads for customers. We saw continued growth in Prime Video ads, which is now available in 16 countries and is contributing meaningfully to our revenue growth. Prime Video has an average ad-supported audience of 315 million viewers globally, up from 200 million in early 2024.

Our ads team is also innovating with AI. We recently announced our ads agent, which lets brands use AI to create and optimize campaigns at scale, implement effective campaign targeting and quickly create actionable insights. And our creative agent lets advertisers research, brainstorm and generate full funnel ad campaigns from concept to completion using conversational guidance in Amazon's retail data, transforming what was a week-long process into just hours.

We're also continuing to invent and see momentum in several other areas, and I'll mention just a few. Starting with live sports on Prime. The fourth season of Thursday Night Football broke more records. It was our most watched season ever, averaging more than 15 million viewers, a 16% year-over-year increase and a third consecutive year of double-digit growth. And the Packers versus Bears wild card game was the most streamed NFL game in history with 31.6 million viewers, clearing the prior mark by more than 4 million.

We just made Alexa+ available to all customers in the U.S., free for Prime members and $19.99 a month for non-Prime members. Alexa+ continues to get even better and more capable, and we've added new ways to interact with Alexa, including a new chat experience at alexa.com, a redesigned mobile app and new integrations with third-party devices like Samsung TVs and BMW cars. We've also added new features like the ability to answer a ring doorbell on a customer's behalf and more ways to shop or manage a home. And finally, the team is making rapid progress on Amazon Leo, which will bring connectivity to consumers, enterprises and governments in places where they don't have broadband connectivity. Our enterprise-grade customer terminal, Leo Ultra, is the fastest satellite Internet antenna ever built, delivering simultaneous download speeds of up to 1 gigabit per second and upload speeds of up to 400 megabits per second. Leo will offer enterprise-grade performance and advanced encryption with secure private networking that bypasses public Internet connecting directly to AWS. We've launched 180 satellites, have more than 20 launches planned in 2026, more than 30 in 2027 and expect to launch commercially in 2026.

We have dozens of commercial agreements already signed, including with AT&T, DIRECTV Latin America, JetBlue and Australia's national broadband network and have many more on the way. It's been an action-packed year of innovation and progress, and we've hit the ground running in 2026.

With that, I'll turn it over to Brian for a financial update.

Thanks, Andy. Starting with our top line financial results. Worldwide revenue was $213.4 billion, a 12% increase year-over-year, excluding the 150 basis point favorable impact of foreign exchange. In Q4, we reported worldwide operating income of $25 billion.

This operating income includes three special charges, which reduced operating income by $2.4 billion. The first charge of $1.1 billion is for the resolution of tax disputes associated with our stores business in Italy and the settlement of a lawsuit. This charge primarily impacts our International segment and is largely recorded in the fulfillment and other operating expense line items. Second charge of $730 million is for the estimated severance costs. This charge impacts all three of our segments and is recorded primarily in the fulfillment, sales and marketing and technology and infrastructure expense line items. The third charge of $610 million is for asset impairments, primarily related to physical stores. This charge primarily impacts the North America segment and is recorded in the other operating expense line.

Moving on to our segment results. In the North America segment, fourth quarter revenue was $127.1 billion, an increase of 10% year-over-year. International segment revenue was $50.7 billion, an increase of 11% year-over-year, excluding the impact of foreign exchange. Worldwide paid units grew 12% year-over-year, which was our highest quarterly growth rate in 2025.

The fourth quarter marked a strong finish to the year as we delivered to customers during the peak holiday season. Our sharp pricing, vast selection and record fast delivery speeds resonated with customers. They appreciate the convenience of receiving their items quickly from gifts for family and friends to everyday essentials and perishable groceries. Our millions of global third-party sellers continue to be an important contributor to our broad selection. In Q4, Worldwide third-party seller unit mix was 61%.

We continue to invest in tools and services, including a comprehensive suite of AI tools that help our selling partners manage and grow their businesses.

Shifting to profitability. North America segment operating income was $11.5 billion with an operating margin of 9%, up from an 8% margin in Q4 of 2024. International segment operating income was $1 billion with an operating margin of 2.1%. Excluding the impact of special charges mentioned earlier, International segment operating margins also expanded year-over-year.

We're pleased with the fulfillment network performance throughout the peak season. We made strong progress improving the cost structure of our network over the past few years. In the U.S., our regionalized network is operating at scale, and we continue to make refinements. This regionalization has improved local inventory placement, leading to faster delivery at lower costs.

Last year, U.S. Prime members received over 8 billion items the same or next day, up more than 30% year-over-year, with groceries and everyday essentials making up half of the total items. For the third year in a row, globally, in 2025, we achieved both our fastest ever delivery speeds for Prime members while also reducing our cost to serve. By leveraging our existing U.S. network, we can now deliver perishable groceries to customers in more than 2,300 cities and towns, all with same-day delivery. We saw significant adoption of this service throughout the year. When customers engage with our perishable offering, they demonstrate notably higher monthly spend compared to those who do not shop the category. We also see that customers shopping perishable groceries add 3x more items to their same-day delivery orders. Looking ahead, we see further opportunity to enhance productivity in our global fulfillment network while delivering at faster speeds for customers. We will continue optimizing inventory placement to drive down distance travel, reduce touches per package and improve package consolidation as well as launch robotics and automation to increase efficiency and elevate the customer experience.

Shifting to advertising. Advertising revenue grew 22% in the fourth quarter, and we added over $12 billion of incremental revenue in 2025 alone as our full funnel advertising approach of connecting brands with customers is resonating simplifying the advertiser experience to enable brands to better reach customers wherever they are.

Moving next to our AWS segment. Revenue was $35.6 billion and growth accelerated to 24% year-over-year. We added $2.6 billion in quarter-over-quarter revenue and AWS now has an annualized revenue run rate of $142 billion. This acceleration was driven by both core and AI services as customers continue to modernize their infrastructure and migrate workloads to the cloud.

Our AI offerings continue to resonate with customers, including our agentic capabilities. This growth was helped in part by the more than 1 gigawatt of capacity we added in Q4. In 2025, AWS added more data center capacity than any other company in the world. AWS operating income was $12.5 billion. We're seeing strong top line growth while remaining focused on driving efficiencies across the business. This includes investing in software and process improvements to optimize server capacity, developing a more efficient network using our lower-cost custom networking gear and advancing custom silicon. At the same time, we continue to rapidly develop products and services on behalf of customers. As we've long said, we expect AWS operating margins to fluctuate over time, driven in part by the level of investments we're making at any point in time.

Turning to cash flows. Our full year operating cash flow increased to $139.5 billion in 2025, up 20% year-over-year due primarily to improved operating income and changes in working capital.

Now turning to our Q1 financial guidance. Q1 net sales are expected to be between $173.5 billion and $178.5 billion. This guidance anticipates a favorable impact of approximately 180 basis points from foreign exchange rates. As a reminder, global currencies can fluctuate during the quarter. Q1 operating income is expected to be between $16.5 billion and $21.5 billion.

A few things to mention on the operating income guidance. Within the North America segment, we do expect a year-over-year cost increase of approximately $1 billion related to Amazon Leo. We have more than 20 launches planned in 2026 and more than 30 in 2027, which means we're spending more on launching satellites each year. Select enterprise customers are testing Amazon Leo services now, and we expect a wider commercial rollout later this year. As a reminder, today, we do expense most of these Leo costs as incurred. We expect that later in the year, many of these costs such as satellite manufacturing and launch services will be capitalized.

Within the International segment, we're continuing to invest more in our stores business to enhance the customer experience and to encourage retail demand to move online more quickly. This includes bringing faster delivery options, including Amazon Now, our service, which delivers to customers in 30 minutes or less.

We're also working hard to stay sharp on pricing and seller fees. And there are countries where we've had to be more aggressive to meet or beat competitors' prices. We like these investments because they will delight customers, grow our business, and we believe they will generate long-term positive return on invested capital.

As we enter 2026, I'm energized by our team's strong execution. I want to thank everyone across the company for their hard work on behalf of our customers. We remain focused on driving an even better customer experience, which is the only reliable way to create lasting value for our shareholders.

With that, let's move on to your questions.

[Operator Instructions] Our first question comes from the line of Mark Mahaney with Evercore ISI.

I think, Brian, let me throw this to you or maybe to Andy. On the strong long-term return on invested capital, I think that's the debate in the market today. So could you give us a little bit more insight into how you think investors will be able to see that, either talk about the duration of the CapEx cycle that you're going through now or what we should see in terms of profitability levels?

And maybe also talk about like other de minimis or minimum free cash flow generation levels that you don't want to go below as you go through this CapEx cycle. Just help us get to that -- get to your level of confidence in having a strong long-term return on that invested capital.

Yes. Sure, Mark. Thank you. I'll start from the financial side. So on the investments we're making, as Andy said earlier, we are putting into service with customers all capacity that we're getting and it's immediately useful. And we're also seeing a long arc of additional revenue that we see from other customers and backlog and commitments that people are anxious to make with us, especially for AI services. So you can see that's working its way into our P&L, both through CapEx and also through our operating margin in AWS. AWS is 35% operating margin through Q4, up 40 basis points year-over-year.

As we talked about before, that is going to fluctuate over time. It certainly has a headwind from the investments in AI and the depreciation on that CapEx. But we also work very hard to offset that with efficiencies and cost reductions. So we will see how that develops over time. So -- but yes, we see strong return on invested capital. We see strong demand for these services, and we continue to like the investments in this area.

I would add to that. If you look at the capital we're spending and intend to spend this year, it's predominantly in AWS. And some of it is for our core workloads, which are non-AI workloads because they're growing at a faster rate than we anticipated. But most of it is in AI. And we just have a lot of growth and a lot of demand. And when you're growing 24% year-over-year with an annualized revenue run rate of $142 billion, you're growing a lot. And what we're continuing to see is as fast as we install this capacity, this AI capacity, we are monetizing it. And so it's just a very unusual opportunity.

As I've shared a lot of times, I passionately believe that every customer experience that we know of today is going to be reinvented with AI, there are going to be a whole bunch of customer experiences that none of us ever imagined that are going to become the norms of how we all operate every day and what we use.

And I think the other thing is that if you really want to use AI in an expansive way, you need your data in the cloud and you need your applications in the cloud. Those are all big tailwinds pushing people towards the cloud. So we're going to invest aggressively here, and we're going to invest to be the leader in this space as we have been for the last number of years. We have, I think, a fair bit of experience over the years in AWS of forecasting demand signals and doing it in such a way that we don't have a lot of wasted capacity and that we also have enough capacity to serve the demand that's there.

And I think we've also proven with AWS over the years in how we build data centers and how we run them and how we invent in there, if you think about our chips and our hardware and our networking gear and how we've invented in power that this isn't some sort of quixotic top line grab, we have confidence that we -- that these investments will yield strong returns on invested capital. We've done that with our core AWS business. I think that will very much be true here as well.

And I think some of the things that you will see over time in the AI space is you're going to keep seeing all of the inference services, which is going to be the majority of the long-term AI workloads is going to be inference. You're going to see the inference keep getting optimized. You're going to see higher utilization on those services. You'll see prices normalize over a period of time.

And then I think the companies that have not just the excellence in infrastructure, but also the components that give them -- give customers better price performance and give those companies themselves better economics are going to have advantaged financials. And I think if you look, we're already off to a really good start having Trainium underneath the majority of our Bedrock service. And that's not just giving customers better prices, but it also gives us better economics. And so we see that following the same sorts of patterns we saw in the early days of our core AWS investment. I'm very confident we're going to have strong return on invested capital here.

And the next question comes from the line of Doug Anmuth with JPMorgan.

Can you just talk about how Project Rainier is running with Anthropic after its first full quarter? And I think in the release, it talks about 500,000 chips, but a few months ago, you talked about getting to 1 million as well. So if you could clarify that.

And then maybe just to follow up on Mark's question. Are there any financial guardrails or governors in place that we should think about around the spend just in terms of operating income growth or positive free cash flow?

Yes. I'll start with the Trainium piece. We are very excited about the growth that we see in Trainium and the future that we have there. I think if you look at what's happened in the early innings of AI over the first few years, you see a lot of usage, but customers are really thirsty for better price performance. And Trainium has 30% to 40% better price performance than comparable GPU. So it's very compelling to customers. You mentioned Project Rainier, Anthropic is building their next -- they're training their next cloud model on top of Trainium2. And that's what Project Rainier is. So we talked about 500,000 chips there. You'll see that continuing to increase. They're also using a fair bit of Trainium2 for other workloads and their own APIs beyond just Project Rainier. But Trainium is a multibillion dollar annualized run rate business at this point, and it's fully subscribed.

And what you're also seeing is Trainium3, which is the next version of Trainium, which we just started shipping, that's 40% more price performance than Trainium2. And we have -- there is a very substantial amount of interest there. We expect that nearly all of that supply will be committed by somewhere around the middle of this year. And we're just in the process of building Trainium4. There's very substantial interest in Trainium4, which is coming in 2027. And we're already having conversations about Trainium5. So there is a lot of interest in Trainium at this point.

And I think when you -- I think people know about our chips capability and our chips business, but I'm not sure folks realize how strong a chips company we've become over the last 10 years. If you look at what we've done with Trainium, if you look at what we've done with Graviton, which is our CPU chip, which is about 40% better price performance than comparable x86 processors, 90% of the top 1,000 AWS customers are using Graviton very expansively. If you combine Trainium and Graviton, it's well over a $10 billion annualized run rate business, and it's still very early there.

So I'm very optimistic about what we're seeing. The Project Rainier has gone very well. I think Anthropic is quite pleased with it. We've learned a lot in the process as well, but it's early days with what's possible here. This is a big business that's getting bigger and has a lot of potential.

And then I just -- I'd briefly comment on your second question that we are as I mentioned, this is what -- I think this is an extraordinarily unusual opportunity to forever change the size of AWS and Amazon as a whole. I think it also is an extraordinary opportunity for companies to change all their customer experiences and for start-ups to be able to build brand-new experiences and businesses that would have taken much longer to try to accomplish before that they can do right now. And so we see this as an unusual opportunity, and we are going to invest aggressively here to be the leaders because like we've been in the last number of years and like I think we will be moving forward.

The next question comes from the line of Ross Sandler with Barclays.

Great. Andy, you mentioned a few calls back how the AI market was currently a bit top heavy with a lot of the spend kind of clustering around a few of the AI native labs. So how is that changing as you look out into '26? And specifically, how do you think you might extend your relationship with a company like OpenAI to maybe help Amazon's AI efforts, both on the retail side and the AWS side?

Yes. The way I would describe what we see right now in the AI space is it's really kind of a barbelled market demand where on one end, you have the AI labs who are spending gobs and gobs of compute right now, along with what I would consider a couple of runaway applications. And then at the other side of the barbell, you've got a lot of enterprises who are getting value out of AI in doing productivity and cost avoidance types of workloads. These are things like customer service or business process automation or some of the fraud pieces. And then in that middle of the barbell are all the enterprise production workloads. And I would say that the enterprises are in various stages at this point of evaluating how to move those, working on moving those and then putting them into production. But I think that middle part of the barbell very well may end up being the largest and the most durable. And I would put in the middle of that barbell, too, by the way, I would put just the altogether brand-new businesses and applications that companies build that right from the get-go run in production on top of AI.

And so I think that to me, when I look at this and what's happening, it's kind of unbelievable if you look at the demand of what you're seeing already with AI, but the lion's share of that demand is still yet to come in the middle of that barbell. And that will come over time. It will come as you have more and more companies with AI talent as more and more people get educated with the AI background. as inference continues to get less expensive, and that's a big piece of what we're trying to do with Trainium and our hardware strategy. And as companies start to have success in moving those workloads to -- further and further success in moving those workloads to run on top of AI. So I think there's -- it's just a huge opportunity. It's still in the relative early stages, even though it's growing at a very -- like an unprecedented clip as we've talked about.

And then I think how do we see our relationships extending with other companies like OpenAI, I would tell you that this movement in what's happening in AI is -- it's very broad. It's going to be a lot of companies. It is a lot of companies already. There's a number of AI labs, but almost every company you talk to, almost every conversation we have on the AWS side, starts with AI. And so we have very significant relationships with a lot of different companies. I think we announced an agreement with OpenAI in November. We're excited about that agreement. It's a big one. We have a lot of respect for the company, and we hope to continue to extend our partnership over time. But this AI movement is not going to be a couple of companies. It's going to be thousands of companies over time.

The next question comes from the line of Michael Morton with MoffettNathanson.

This one is on the retail business. Andy, you've talked about how you're passionate this is going to change experiences across the board. And you've shared some encouraging data points on Rufus. And we're seeing all of the other Internet platforms roll out Agentic protocols. I would love to see how you think this plays out for the retail business and the on-site ads portion of the retail business is what seems like it could be a compression in the funnel as consumers get better answers over time. Anything there would be great.

I'm very optimistic about the customer experience that will ultimately be what customers use for Agentic shopping. And I think it's good for customers. I think it's going to make it easier for them. It's a big piece of why we've invested as significantly as we have in our own shopping assistant in Rufus. And if you haven't checked out Rufus recently, I really encourage you to do so. It's gotten much, much better and keeps getting better every month. And we have about -- we have 300 million customers who used Rufus in 2025. Customers who use Rufus are about 60% more likely to complete a purchase. And so you just -- you're seeing a lot of usage of it and a lot of growth, and I think it's very useful.

And I think at the same time, we will have relationships with third-party horizontal agents that can enable shopping as well. We have to collectively figure out a better customer experience. It's still -- these horizontal agents don't have any of your shopping history. They get a lot of the product details wrong, they get a lot of the pricing wrong. And so we have to try to find a customer experience together that's better and a value exchange that makes sense for both parties. But I'm very hopeful that we'll get there over time. We continue to have a number of conversations.

And then I think you're going to have to look at as time goes on, which types of -- which shopping agents are consumers going to use. And it kind of reminds me in some ways of the early days of kind of all the search engines that were referring traffic to retailers. And it's still a relatively small portion of the overall traffic and sales. But of that fraction, you have to ask how many consumers are going to prefer using a horizontal agent where it's kind of a middle person between the retailer and the consumer versus wanting to use a great agent from that retailer that has all its shopping history and that has all the data right there and makes it easy if you're just spearfishing for something to shop for it right there or if you want to do discovery, you can do it there, and it's got the best data on shopping.

I think a lot of customers are ultimately going to choose to use a great shopping agent from that retailer. Because if you think about what consumers really want in retail in a retailer, they want really broad selection. They want low prices. They want really fast delivery. And then they want a retailer that they can trust and that takes care of them. And I think horizontal agents are pretty good at aggregating selection, but retailers are much better at doing all 4 of those items. And so I'm very optimistic that people will use our shopping agent. It's off to a great start. I also expect that we'll work with other third-party agents over time as we work on the issues I mentioned earlier.

The next question comes from the line of Brian Nowak with Morgan Stanley.

Andy, I want to ask you one about the global retail business this year. I know there's a lot of areas of investment in it that you're talking about to sort of make -- improve the service, make it more durable over the long term, et cetera. But I'm assuming there are also sources of efficiency you expect to see this year. So can you sort of help us understand both sides of the ledger on retail this year? Where are some of the areas where you see the potential for sources of efficiency and cost to serve savings? And then where should we be thinking about the areas of investment to sort of drive more durable growth, robotics, et cetera? How does that sort of break down?

Yes. So I would say on the side of continuing to invest to keep growing the retail business, the kind of core drivers of demand continue to be the same. We're going to work really hard to expand selection. And you've seen what we've done over the last several years. The expansion of selection has been broad. And you'll see it on both ends of the spectrum. We have a lot more of those luxury brands that have built presences in Amazon had success and found that we could manage their brand presentation in the right way, and they've been very happy. I mean you only have to look at L'Oreal as an example, too, of just how fast that business is growing and how happy our partners have been. And at the same time, we are working really hard to continue to expand the amount of everyday essentials that we offer our customers.

And the growth in Everyday Essentials in our business is really remarkable, as I mentioned in my opening comments. And one out of three units now that we move our Everyday Essentials. And what we find there is that the more the customers can rely on us for Everyday Essentials and the lower ASP items, they just choose to do more of their downstream shopping with us in every way. We're just more front of mind.

And so I think a big piece of why we have captured more and more of those everyday essentials, and you see it also in our grocery business with perishables, too, is just our speed of delivery improvements over the last three years has been really marked. I mean it's customers -- it's the one thing I get stopped on the street most often about, which is I just can't believe how quickly from when I order something, I get it to my door and how reliable you are.

I think along that speed of delivery piece, it's also quite interesting what's happening with quick commerce, and we have this offering called Amazon -- now that we've largely started outside the U.S. in India and the UAE and Mexico that gets thousands of items to customers within 30 minutes. And it really is -- it's quite interesting how quickly that is growing. And I think that it's just another one of those things like Everyday Essentials that when you're able to order more and more from Amazon, you just think of Amazon first if it's a great experience that we're offering for whatever you're looking for. But in our -- if you look in India, which is the place we've rolled out quick commerce the fastest, customers who try quick commerce are shopping with triple the frequency than they did before they tried us in quick commerce.

So, those are all areas, I think, are pretty excited that we're expanding. You'll see us continue to expand what we're doing on the perishable side, too, which we're quite excited about. And we are able to deliver perishables same day in thousands of cities around the world now. And the cities in which we have those perishables available, 9 of the 10 top items that are ordered in that geography are perishables. And so we're just having a lot of success with that, too. And people buy perishables from us after they buy perishables, they're shopping with us twice as frequently. So a lot of good things to like there.

And then on the efficiencies, we are -- I mean, we always have a very long list of these that we're working on, Brian. And it's true today as well. Like if you look -- even -- I mentioned -- I talked a lot about regionalization in our fulfillment network, particularly in the U.S. over the last couple of years. And I said we weren't done honing that, and that's true. It's just -- we don't talk about it every time. But if you look at what we've done there, we've extended the number of regions. It was 8. It's now 10. We've extended regionalization to what we do with our inbound delivery to be much more efficient in being able to get more items closer to customers more quickly.

We have made a lot of -- we're doing a lot of work, and we've made a huge amount of progress in being able to get more units into each box. And as we're able to get more units into each box, it obviously saves shipments and we drive better operating income when we do that. And we've made very significant progress there, but have a lot more planned. It's part of -- by the way, that improvement is part of what helps us do things like I was talking about earlier and adding to a delivery in near real time.

And then robotics, as you mentioned, is another big one for us. We have over 1 million robots today in our fulfillment network. They take care of all sorts of functions, but still a fraction of what I think we're going to be able to enable over time, which will allow our -- we'll always have a lot of people that we employ in our fulfillment network, but they'll leave to the robotics things that are more repetitive. So it's better productivity for the business, more safe for our teammates and there's real cost efficiencies in that as well. So a lot on both sides of the ledger as always.

And our final question comes from the line of Eric Sheridan with Goldman Sachs.

Maybe a few parts just on AWS. Can you speak to the current state of your revenue backlog as of Q4 and also discuss a little bit about what you see both for internal use cases and external client needs with respect to any imbalance between supply and demand around AI efforts and how you think about closing the gap on those as more capacity comes online through 2026?

That's a lot of parts. I'll start with the first one, which is on backlog, our backlog is $244 billion. That's up 40% year-over-year. I think it's up 22% quarter-over-quarter. We have -- and we have a lot of deals that are in the pipeline. There's just -- as I mentioned earlier, there is a lot of demand for AWS right now. in the AI space and also in the core AWS space.

Your second question was internal and external use cases. and then the impact around supply and demand. The vast majority of our -- the capital that we spend and the capacity that we have is consumed by external customers. We have -- Amazon has always been a very large AWS customer, a very helpful AWS customer because they're very demanding, and they use the services very expansively and stretch the limits as we launch things. So they've always been a very important big customer, but always a very small fraction of the total, and that's true today in AI as well as the overall AWS business.

Internally, we have all sorts of ways that we are using AI. We have over 1,000 AI applications that we've either deployed or in the process of building, and they range from our shopping assistant in Rufus that we were just talking about to Alexa+, which is a really large-scale generative AI application to applications in our fulfillment network that allow us to have more accurate forecasting predictions to how we do customer service and our customer service chatbot to how we are making it much easier for brands to create advertisements and to optimize all their campaigns across the full funnel of advertising options we have to -- in live sports, if you watch Thursday Night Football, you can see defensive alerts, which predict which player is going to blitz or pocket health. I mean we -- in every one of our businesses, you see a very broad use of AI to improve the customer experience. And in many cases, just to completely reinvent what was possible before. I mean it's pretty neat to use something like Lens where you may see something you want to buy, you can just take a picture of it in the app and it finds the item on the detail page you can buy in one click. It's kind of magic.

And externally, I would say it's kind of what I said earlier. You have AI labs consuming lots and lots of capacity. both for training as well as for the inference and the research across what they're doing with their different applications and models. We see enterprises all sorts of workloads, customer service automation, business process automation, fraud, completely reinventing their applications, agentic coding applications, legal applications. Suno is a really cool example of an AWS customer that's kind of reinvented how you can write music and build music. So really across the board.

And I just think on the supply and demand, what I would tell you is we're growing 24% year-over-year on a $142 billion annualized run rate business. So we're growing at really an unprecedented rate yet, I think every provider would tell you, including us that we could actually grow faster if we had all the supply that we could take. And so we are being incredibly scrappy around that.

If you look in the last 12 months, we added 3.9 gigawatts of power. Just for perspective, that's twice what we had in 2022 when we were an $80 billion annual run rate business. We expect to double it again by the end of '27. We added 1.2 gigawatts of power in Q4, just quarter-over-quarter. So it's -- so we are -- our team is being aggressive and scrappy and inventive in adding capacity as fast as we can. We'll add a lot more in '26 and '27 and '28 for that matter. And -- and we're very optimistic we can continue to grow in the ballpark of what we have.

Thanks for joining us on the call today and for your questions. A replay will be available on our Investor Relations website for at least three months. We appreciate your interest in Amazon, and we look forward to talking with you again next quarter.

Thank you for standing by. Good day, everyone, and welcome to the Amazon.com Third Quarter 2025 Financial Results Teleconference. [Operator Instructions] Today's call is being recorded.

And for opening remarks, I will be turning the call over to the Vice President of Investor Relations, Mr. Dave Fildes. Thank you, sir. Please go ahead.

Hello, and welcome to our Q3 2025 financial results conference call. Joining us today to answer your questions is Andy Jassy, our CEO; and Brian Olsavsky, our CFO. As you listen to today's conference call, we encourage you to have our press release in front of you, which includes our financial results as well as metrics and commentary on the quarter. Please note, unless otherwise stated. All comparisons in this call will be against our results for the comparable period of 2024. Our comments and responses to your questions reflect management's views as of today, October 30, 2025 only, and will include forward-looking statements.

Actual results may differ materially. Additional information about factors that could potentially impact our financial results is included in today's press release and our filings with the SEC, including our most recent annual report on Form 10-K and subsequent filings.

During this call, we may discuss certain non-GAAP financial measures. In our press release, slides accompanying this webcast and our filings with the SEC, each of which is posted on our IR website. You will find additional disclosures regarding these non-GAAP measures, including reconciliations of these measures with comparable GAAP measures.

Our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our results are inherently unpredictable and may be materially affected by many factors, including fluctuations in foreign exchange rates, changes in global economic and geopolitical conditions, tariff and trade policies and customer demand and spending, including the impact of recessionary fears; inflation, interest rates, regional labor market constraints, world events, the rate of growth of the Internet, online commerce cloud services and new and emerging technologies and the various factors detailed in our filings with the SEC. Our guidance assumes, among other things, that we don't conclude any additional business acquisitions, restructurings or legal settlements. It's not possible to accurately predict demand for our goods and services, and therefore, our actual results could differ materially from our guidance.

And now I'll turn the call over to Andy.

Thanks, Dave. We saw strong growth across our business in Q3, and we're reporting $180.2 billion in revenue, up 12% year-over-year, excluding the impact from foreign exchange rates. Operating income was $17.4 billion, but would have been over $21 billion, if not for 2 special Q3 expenses, $2.5 billion for an FTC settlement and $1.8 billion for estimated severance costs. Trailing 12-month free cash flow was $14.8 billion.

I'll start with AWS. AWS is growing at a pace we haven't seen since 2022, reaccelerating to 20.2% year-over-year, our largest growth rate in 11 quarters. It's worth remembering that year-over-year percentage growth is a relative term. It's very different having 20% year-over-year growth on a $132 billion annualized run rate and to have a higher percentage growth rate on a meaningfully smaller annual revenue, which is the case with our competitors. Backlog grew to $200 billion by Q3 quarter end and doesn't include several unannounced new deals in October, which together or more than our total deal volume for all of Q3. AWS is gaining momentum. Customers want to be running their core and AI workloads in AWS given its stronger functionality, security and operational performance and the scale I see in front of us gives me significant confidence in what lies ahead. I'll share a little more detail on why.

It starts with AWS having much broader infrastructure functionality. Start-ups, enterprises and governments want to move their production workloads to the place that has the broadest and deepest array of capabilities. AWS has more services and deeper features within those services than anybody else and continues to innovate at a rapid clip. These are key building blocks for anything that customers want to create, and they're a big part of why Gartner has named AWS leader in its strategic cloud platform services Magic Quadrant for 15 consecutive years. We're bringing the same building block approach to AI. SageMaker makes it much simpler for companies to build and deploy their own foundation models. Bedrock gives customers leading selection of foundation models and superior price performance to deploy inference into their next-generation applications. A lot of the future value companies will get from AI will be in the form of agents. AWS is heavily investing in this area and well positioned to be a leader.

Companies will both create their own agents and use agents from other companies. For those building their own, it's been harder to build than it should be. It's why we launched strands to make it much easier to create agents from any foundation model that builders desire. For companies who successfully built agents, they've hesitated putting them into production because they lack secure scalable runtime services or memory or observability built specifically for agents. It's why we launched AgentCore, a set of infrastructure building blocks that allow builders to deploy secure, scalable agents. Ericsson used AgentCore to deliver AI agents across their workforce, Sony used it to build a agentic AI platform with enterprise-level security, observability and scalability. And Cohere Health is using AgentCore to deploy agents that will reduce medical review times by up to 30% to 40%. AgentCore's SDK has already been downloaded over 1 million times, and our builders are excited about it. It's an enabler.

Companies will also use other agents, and AWS continues to build many of the agents we believe builders will use in the future. For coding, we've recently opened up our agentic coding IDE called Kiro. More than 100,000 developers jumped into Kiro in just the first few days of preview and that number has more than doubled since. It's processed trillions of tokens thus far, weekly actives are growing fast, and developers love its unique spec and tool call and capabilities. For migration and transformation, we offer an agent called Transform. Year-to-date, customers have already used it to save 700,000 hours of manual effort. The equivalent of 335 developer years of work. For example, Thomson Reuters used it to transform 1.5 million lines of code per month, moving from Windows to open source alternatives and completing tasks or a times faster than with other migration tools.

Customers have also already used Transform to analyze nearly 1 billion lines of mainframe code as they move mainframe applications to the cloud. For business customers, we've recently launched QuickSleep to bring a consumer AI-like experience to work, making it easy to find insights, conduct deep research, automate tasks, visualize data and take actions. We've already seen users churn months long projects in today's get 80% plus time savings on complex tasks and realize 90% plus cost savings. And for contact centers, we offer Amazon Connect which creates a more personalized and efficient experience for contact center agents, managers and their customers. Connect has recently crested $1 billion annualized revenue run rate with 12 billion minutes of customer interactions being handled by AI in the last year and is being used by large enterprises like Capital One, Toyota, American Airlines and Ryanair. These are real practical results for customers, and there are many more examples like them.

Because of its advantaged capabilities, security, operational performance and customer focus, AWS continues to earn most of the big enterprise and government transformations to the cloud. As a result, AWS is where the preponderance of company's data and workloads reside and part of why most companies want to run AI and AWS. To enable customers to do so, we need to have the requisite capacity, and we've been focused on accelerating capacity the last several months, adding more than 3.8 gigawatts of power in the past 12 months, more than any other cloud provider. To put that into perspective, we're now double the power capacity that AWS was in 2022, and we're on track to double again by 2027. In the last quarter of this year alone, we expect to add at least another 1 gigawatt of power. This capacity consists of power, data center and chips, primarily our custom silicon, Trainium and NVIDIA.

We've recently brought Project Rainier Online, our massive AI compute cluster spanning multiple U.S. data centers and containing nearly 500,000 of our Trainium2 chips. Anthropic is using it now to build and deploy its industry-leading AI model, Claude, which we expect to be on more than 1 million Trainium2 chips by year-end. Trainium2 continues to see strong adoption, is fully subscribed is now a multibillion-dollar business that grew 150% quarter-over-quarter. Today, Trainium is being used by a small number of very large customers but we expect to accommodate more customers starting with Trainium3. We're building Bedrock to be the biggest inference engine in the world and in the long run, believe Bedrock could be as big a business for AWS as EC2, and the majority of token usage in Amazon Bedrock is already running on Trainium. We're also continuing to work closely with chip partners like NVIDIA, with whom we continue to order very significant amounts as well as with AMD and Intel. These are very important partners with whom we expect to keep growing our relationships over time.

You're going to see us continue to be very aggressive in investing in capacity because we see the demand. As fast as we're adding capacity right now, we're monetizing it. It's still quite early and represents an unusual opportunity for customers in AWS.

I'll now turn to stores. Where the team continues to deliver and innovate for customers across our key priorities, selection, low prices and convenience, particularly fast delivery, we're offering 14% more selection since last quarter from popular brands like The North Face and Charlotte Tilbury, and we've added hundreds of thousands of items from popular brands this year. Everyday Essentials continues to grow quickly, and year-to-date is growing nearly twice as fast as the rest of the business. We continue to make it easier for customers to order low-priced perishable groceries from Amazon, and customers in more than 1,000 cities and towns now can shop fresh groceries alongside millions of Amazon.com products with free same-day delivery. This is a game changer for customers who can now order milk alongside electronics, check out with one cart and have everything delivered to their doorstep within hours.

The team also invented a new add to delivery button that lets customers add items to previously scheduled orders and it's been used more than 80 million times since launch, and it's just launch. It's an example of one of those seemingly simple but powerful innovations that make customers' lives easier. We remain committed to staying sharp on price and meeting or beating prices of other major retailers. In July, we had our biggest Prime Day event ever, with customers saving billions of dollars across more than 35 categories. We continue to break records on speed. We're on track to deliver at our fastest speeds ever for Prime members globally once again this year, and we've started rolling out 3-hour delivery in select U.S. cities.

We're also continuing to invest in infrastructure to speed up rural deliveries and serve more customers in more communities. That includes committing over $4 billion to expand our rural delivery network across the U.S. These are small towns where people want fast delivery, but where other companies have been backing out and reducing service. In contrast, we've already increased the number of rural communities with access to our same-day and next-day delivery by 60%, reaching roughly half of the total communities we plan to expand to by the end of the year. The stores team is also innovating rapidly with AI. For example, Rufus, our AI-powered shopping assistant has had 250 million active customers this year with monthly users up 140% year-over-year, interactions up 210% year-over-year and customers using Rufus during a shopping trip being 60% more likely to complete a purchase. Rufus is on track to deliver over $10 billion in incremental annualized sales.

Here are the highlights. Our generative AI-powered audio feature that combines product summaries and reviews to make shopping easier has expanded from hundreds of products at launch to millions of products and millions of customers have used it streaming almost 3 million minutes. In Amazon Lens, an AI-powered visual search tool that lets customers find products with their phones camera, a screenshot or a bar code, now includes Lens Live, which instantly scans products and shows real-time matches in a swipeable carousel. Tens and millions of customers are using Amazon Lens each month.

Moving on to Amazon ads. We're pleased with the continued strong growth, generating $17.6 billion of revenue in the quarter and growing 22% year-over-year. We see strength across our broad portfolio of full photo advertising offerings that helps advertisers reach an average ad-supported audience of more than $300 million in the U.S. alone. We also continue to be excited about our demand side platform, Amazon DSP, which lets advertisers plan, activate and measure full funnel investments. Last quarter, I mentioned our partnership with Roku and we've built on that with a partnership with Netflix, providing advertisers using Amazon DSP with direct access to Netflix's premium ad inventory. We announced integrations with Spotify and SiriusXM. With Spotify, we provide advertisers with direct programmatic access to a global audience of more than 400 million monthly ad-supported listeners. And with SiriusXM, brands can reach 160 million monthly digital listeners across services like Pandora and SoundCloud and we're excited about the advertising opportunity around prime video live sports. Live sports got a lot of interest from advertisers in upfront negotiations for 2025, '26, and we exceeded our own expectations for upfront commitments with significant growth across the board.

Finally, we're continuing to invade for advertisers with AI. For example, in September, we announced an agentic AI tool and creative studio that plans and executes the entire creative process in a matter of hours instead of weeks. We're also inventing and seeing strong momentum in several other areas, and I'll mention just a few. In Prime Video live sports, NBA on Prime tipped off last week and our opening night doubleheader averaged 1.25 million viewers in the U.S., a double-digit increase over last season on cable. You'll see us bring the same constant innovation here that we brought to our NFL broadcast. We're adding golf with The Masters in 2026 and new skins competition with the PGA Tour on Black Friday this year. And we've added Peacock and FOX One to Prime Videos add-on subscription offering of over 100 channels in the U.S.

We continue to be energized by the response to Alexa+ compared to what we call the classic Alexa experience, Alexa+ customers are talking to Alexa 2x more. Those interactions are much longer, and they're covering a broader range of topics. So using Alexa+ and Fire TV at 2.5x the rate of classic using natural conversation to discover audio content 4x more, engaging with photos 4x more and customers are completing 4x more shopping conversations that end in a purchase. We've expanded the number of project hyper satellites and space to more than 150 and delivered over 1 gigabit per second speeds and test with our enterprise-grade customer terminal, the first commercial phased array we know of to clear that threshold. Finally, Zoox robotaxis are available to riders in Las Vegas, and we've announced Washington, D.C. as the eighth testing location. We're excited for these to continue rolling out to more riders. Q4 is one of our busiest and most energizing times of the year, and we're excited about the continued demand for AWS. The innovations will announce the reinvent in December, the positive customer response to our AI-powered experiences, all the guests will be delivering throughout the holiday season and a lot more.

Thanks in advance to our teammates around the world who are gearing up to deliver for customers once again. With that, I'll turn it over to Brian for a financial update.

Thanks, Andy. Starting with our top line financial results. Worldwide revenue was $180.2 billion, a 12% increase year-over-year, excluding a 90 basis point favorable impact of foreign exchange. In Q3, we reported worldwide operating income of $17.4 billion. This operating income includes 2 special charges, which reduced operating income by $4.3 billion. The first charge of $2.5 billion is related to a legal settlement with the Federal Trade Commission, which impacts the North America segment and is recorded in the other operating expense line.

The second charge of $1.8 billion relates to severance costs for roll eliminations and impacts all 3 of our segments. The severance charge is recorded primarily in the technology and infrastructure, sales and marketing and general and administrative expense line items. Excluding these 2 charges, worldwide operating income would have been $21.7 billion or $1.2 billion above the high end of our guidance range.

Moving to our segment results. We remain encouraged by the innovation our teams are delivering for customers across all 3 segments. In the North America segment, third quarter revenue was $106.3 billion, an increase of 11% year-over-year. International segment revenue was $40.9 billion, an increase of 10% year-over-year, excluding the impact of foreign exchange. Worldwide paid units grew 11% year-over-year. We continue to prioritize the inputs that matter most to our customers. In the third quarter, our sharp pricing, broad selection and fast delivery speeds continue to resonate with customers. Customers appreciate the ability to quickly receive items essential for their daily needs, including perishable groceries and have them delivered in the same day. Our millions of global third-party sellers continue to be important contributors to our vast selection, which helps customers find the items they need at competitive prices.

We're committed to building innovative services and features for our sellers, including our ongoing advancements in generative AI. Today, more than 1.3 million sellers have used our generative AI capabilities to more quickly launch high-quality listings. Better listings translate into better traction with customers. And in Q3, worldwide third-party seller unit mix was 62%, up 200 basis points from Q3 of last year. Shifting to profitability. North America segment operating income was $4.8 billion, with an operating margin of 4.5%. Excluding the $2.5 billion charge related to the legal settlement with the FTC, North America segment operating income would have been $7.3 billion with an operating margin of 6.9%. North America segment operating margin also includes a portion of the severance charge. International segment operating income was $1.2 billion, with an operating margin of 2.9%. Excluding the impact of the severance charge International segment operating margins expanded year-over-year.

Globally, our progress on key inputs is delivering a better customer experience while driving a more efficient cost structure. For example, we're making notable strides in improving inventory placement to speed up delivery to customers. And as a result, for the third year in a row, we are on track to deliver our fastest speeds ever for Prime members in 2025. We continue to tune and improve our fulfillment operations and our regionalized network is operating at scale. We see many benefits of our inbound process improvements, including a reduction of U.S. inbound lead time by nearly 4 days compared to last year. This allows us to be more efficient with our inventory purchasing, which benefits working capital. We're also placing inventory more strategically throughout the network. And by leveraging our existing infrastructure, we're now offering U.S. customers the ability to order perishable groceries and receive them the same day and as little as 5 hours.

We're seeing positive early results since launching in January, when customers start shopping groceries on Amazon, they are visiting the site more often and returning twice as often as nonperishable shoppers. Looking ahead, we see further opportunity to improve our activity in our global fulfillment and transportation network. We continue to improve inventory placement to drive down distance travel and touches for package. We will also build on the gains from our regionalized network through algorithmic improvements as well as launching robotics and automation. Operating margin may fluctuate quarter-to-quarter, we have a delivered approach to achieve sustained progress over the long term.

Shifting to advertising. Advertising revenue was $17.7 billion and growth accelerated for the third consecutive quarter. We continue to see strong growth on an increasingly large base, as our full funnel advertising approach of connecting brands with customers is resonating.

Moving next to our AWS segment. Revenue was $33 billion, up 20.2% year-over-year. This is an acceleration of 270 basis points compared to last quarter, driven by strong growth across both our AI and core services and more capacity, which has come online to support customer demand. AWS revenue increased $2.1 billion quarter-over-quarter and now has an annualized revenue run rate of $132 billion. AWS operating income was $11.4 billion, and reflects our continued growth, coupled with our focus on driving efficiencies across the business. We are expanding our data center footprint, largely to accommodate Gen AI. And to the extent those assets were placed into service related to depreciation does impact our margins. As we've long said, we expect AWS operating margins to fluctuate over time, driven in part by the level of investments we're making at any point in time.

Now turning to our cash CapEx, which was $34.2 billion in Q3. We've now spent $89.9 billion so far this year. This primarily relates to AWS as we invest to support demand for our AI and core services and in custom silicon, like Trainium as well as tech infrastructure to support our North America and international segments. We'll continue to make significant investments, especially in AI, as we believe it to be a massive opportunity with the potential for strong returns on invested capital over the long term. Additionally, we continue to invest in our fulfillment and transportation network to support the growth of the business, improve delivery speeds and lower our cost to serve. These investments will support growth for many years to come. Looking ahead, we expect our full year cash CapEx to be approximately $125 billion in 2025, and we expect that amount will increase in 2026.

I'll finish up my remarks with net income. While we primarily focus our comments on operating income, our third quarter net income of $21.2 billion includes a pretax gain of $9.5 billion related to our investment in Anthropic. This investment activity is not related to Amazon's ongoing operations and is included in nonoperating income. We're encouraged by the start of the peak season and we are ready to serve customers in the coming months. I want to thank our teams across Amazon for their hard work as we get ready to delight customers during the holiday season. Our commitment to elevating the customer experience is the only reliable way to drive sustainable value for our shareholders.

With that, let's move on to your questions.

[Operator Instructions] And our first question comes from the line of Justin Post with Bank of America.

I'll ask on AWS. Can you just kind of go through how you're feeling about your capacity levels and how capacity constrained you are right now? And then in your prepared remarks, you mentioned Trainium3 demand and maybe broadening out your customer base. Can you talk about the demand you're seeing outside of your major customers for Trainium?

Yes. On the capacity side, we brought in quite a bit of capacity, as I mentioned in my opening comments, 3.8 gigawatts of capacity in the last year with another gigawatt plus coming in the fourth quarter and we expect to double our overall capacity by the end of 2027. So we're bringing in quite a bit of capacity today, overall in the industry, maybe the bottleneck is power. I think at some point, it may move to chips, but we're bringing in quite a bit of capacity. And as fast as we're bringing in right now, we are monetizing it.

And then on the Trainium demand, outside of our major customers. So first of all, as I mentioned on Trainium2, it's really doing well. It's fully subscribed on Trainium2. We have -- it's a multibillion-dollar business at this point. It grew 150% quarter-over-quarter in revenue. And you see really big projects at scale now, like our Project Rainier that we're doing with Anthropic, where they're running their next version of -- they're training the next version of Claude on top of Trainium2 on 500,000 Trainium2 chips going to 1 million Trainium2 chips by the end of the year. As I mentioned, we have -- today, with Trainium2, we have a small number of very large customers on it. But because Trainium is 30% to 40% more price performance than other options out there, and because as customers, as they start to contemplate broader scale of their production workloads, moving to being AI-focused and using inference, they badly care about price performance. And so we have a lot of demand for Trainium. Trainium3 should preview at the end of this year with much fuller volumes coming in the beginning of '26, we have a lot of customers, both very large, and I'll call it, medium-sized who're quite interested in Trainium3.

And the next question comes from the line of Brian Nowak with Morgan Stanley.

Congrats on the quarter, guys. So maybe 2. One, Andy, sort of a philosophical chip question. There's a lot of questions in the market about Trainium and sort of its positioning versus other third-party chips. So how do you think about the key hurdles of Trainium3 need to overcome to really make Trainium adoption broader, to your point on the last question and continue to drive Trainium as opposed to satisfying what could be broader demand with third-party chips in the near term?

Yes. Well, first of all, we're always going to have multiple chip options for our customers. It's been true in every major technology building block or component that we've had in AWS. Really in the history of AWS, it's never just one player that over a long period of time has the entire market segment and then can satisfy everybody's needs on every dimension. And so we have a very deep relationship with NVIDIA. We have for a very long time. And we will for as long as I can foresee the future. We buy a lot of NVIDIA. We are not constrained in any way in buying NVIDIA, and I expect that we'll continue to buy more NVIDIA both next year and in the future.

But we're different from most technology companies in that we have our own very strong chip team, and this is our Annapurna team. And you saw it first on the CPU side with what we built with Graviton which is about 40% better price performance than the other x86 processors, and you're seeing it again on the custom silicon on the AI side with Trainium, which is about the same amount of price performance benefit for customers relative to other GPU options. And our customers to be able to use AI as expansively as they want. And remember, it's still relatively early days at this point. They're going to need better price performance and they care about it deeply. And so I mentioned earlier the momentum that Trainium2 has. And I think that for us, as we think about Trainium3, I expect Trainium3 will be about 40% better than Trainium2 and Trainium2 is already very advantaged on price performance.

So we have to, of course, deliver the chip. We have to deliver it in volumes and deliver it quickly. And we have to continue to work on the software ecosystem, which gets better all the time. And as we have more proof points like we have with Project Rainier with what Anthropic's doing on Trainium2, it builds increasing credibility for Trainium. And I think customers are very bullish about it. I'm bullish about it as well.

And our next question comes from the line of Doug Anmuth with JPMorgan.

I'll stick with basically the same topic, Andy. But can you just talk a little bit about the architecture of Project Rainier and how it's differentiated and what that means for customers and for AWS? And do you expect Rainier to expand beyond Anthropic? And how do you replicate Rainier with Trainium3 chips?

Yes. I think what is compelling for entropic around Project Rainier is really is the Trainium2 chip, which we built a very -- first of all, we built a very large cluster that they can use in a very expansive way. And it's not simple to be able to build a cluster that has 500,000 plus chips going to 1 million. That's an infrastructure feet that's hard to do at scale. And so some piece of it is the infrastructure capabilities that we've built over a long period of time in AWS that is unusual in the industry. But it's just also the performance of the chip and the price performance, both of which matter.

And I think that Project Rainier is something that is specific for anthropic, but we have a lot of other customers who are interested in employing large clusters of Trainium chips that we're going to hopefully give them a chance to do so with Trainium3.

The next question comes from the line of Mark Mahaney with Evercore ISI.

I want to ask about 2 topics, groceries and then how to think about head count in the future. And on groceries, I want to -- the perishables, I think last quarter, you talked about 70% or something of users had never purchased from perishables from Amazon before. Just talk about whether you -- I think you used the term, Game Changer, before. Does this mean that maybe we don't -- you no longer need to Amazon Fresh stores. You always had this DVD delivery van density advantage. And have you kind of reached a point you think of scale and speed that you really can change people habit and really have them consider Amazon as one of their first grocery options? Do you really feel like you're at that point?

And then secondly, just on the head count, some of the recent news. Just talk to us about how you think about head count going forward? Are you seeing -- is the level of efficiencies that you're getting from AI such that you can keep head count relatively flattish for the foreseeable future? Just talk about the pros and cons or the wins and losses in terms of that head count going forward?

Yes. So I'll start with grocery, Mark. We have a very large grocery business. If you look at our entire grocery business, if I don't even count Whole Foods Market and Fresh, in the last 12 months to over $100 billion of gross merchandising sales, which would make us a top 3 grocery in the U.S. A good chunk of it is a lot of the items that you'd find in the middle aisle so consumables and canned goods and pet food and health and beauty, very significant and continues to grow at a very good clip.

But then we also have Whole Foods Market, which is the pioneer in organic foods, which is also growing at a faster clip than most grocery companies with an attractive trajectory on profitability, and we'll expand our Whole Foods physical presence over the coming years here. And I'm also very excited about this new concept, daily shop that we have, which is a smaller version of Whole Foods in urban settings, which we have 3 that we've launched that are off to very good starts that you should expect to see more of as well.

And we have always been -- as you referenced, we've talked a lot about having a larger mass physical presence. And we continue to experiment with various formats. But the one that we are most excited about is what you referenced, which is the ability to provide perishable groceries with same-day deliveries. And if you think about how many of our customers are buying from us multiple times a week and who are buying things like shampoo or detergent or paper cups or water, where the ability to add milk and eggs and yogurt and other perishables to their order and have it live in the same shop in cart and then show up a few hours later, is very compelling. And we started with a few markets about a year ago, and we were really taken aback at the adoption, not just the number of people that started buying perishables from us very quickly but how often they came back downstream to buy perishables and groceries from us in the future. And so we've now expanded that to 1,000 cities around the U.S. and will be in 2,300 by the end of the year. And it's really changing the trajectory and the size of our grocery business.

And I also believe that this many years tradition of the weekly stock up grocery stock up is changing. And I think we're a big part of that. And I think there's a lot of potential there for the grocery side. It doesn't mean that we won't continue to experiment with other physical formats, but we're on to something very significant with what we're doing with perish both from our same-day facilities.

And then on your head count question, what I would tell you is the announcement that we made a few days ago was not really financially driven and it's not even really AI-driven, not right now, at least. It really -- its culture. And if you grow as fast as we did for several years, the size of businesses, the number of people, the number of locations, the types of businesses you're in, you end up with a lot more people than what you had before, and you end up with a lot more layers. And when that happens, sometimes without realizing that you can weaken the ownership of the people that you have who are doing the actual work and who own most of the 2-way door decisions, the ones that should be made quickly and right at the front line, and it can lead to slowing you down. And as a leadership team, we are committed to operating like the world's largest start-up. And that means removing layers. It means increasing the amount of ownership that people have, and it means inventing and moving quickly.

And I don't know if there's ever been a time in the history of Amazon or maybe business in general with the technology transformation happening right now, where it's important to be lean, it's important to be flat, and it's important to move fast, and that's what we're going to do.

And the next question comes from the line of Eric Sheridan with Goldman Sachs.

Wanted to know, Andy, if you could reflect on the opportunity that's continuing to present itself in terms of rolling out more robotics and automation and the broader theme of physical AI across your operations? And how should we be thinking about that as a driver of potential efficiencies, but also as a driver of the ability to possibly reinvest back in the business over the long term?

Robotics is a very substantial area of investment for us. We have over 1 million robots in our fulfillment network at this point. And I would say that while that's significant, we have a lot of invention in flight. So I expect that we'll have more over a period of time.

Robotics are very important for us and for our customers and for our teammates because they improve safety, they boost productivity, they increased speed, and they let our human teammates focused on problem solving and what they do best. And we expect that our people remain at the heart in the center of our fulfillment network as they have from when we first started working the robotics. And we expect that over time, we will have a fulfillment network where robots and humans complement each other and work together. But I think you're going to continue to see us advance invest very significantly in robotics. It's going to help on the safety, the productivity, the speed and ultimately some of the cost pieces, which will allow us to continue to improve the customer experience.

And the next question comes from the line of John Blackledge with TD Cowen.

How does Amazon think about agentic commerce going forward? And how do you think Amazon will serve customers using agents to purchase goods on Amazon in the future?

I'm very excited about -- and as a business, we're very excited about in the long term the prospect of agentic commerce. And it has a chance to be good for customers has a chance to be really good for e-commerce. And I think if you're -- if you know what you want to buy, there are a few experiences that are better than coming to Amazon. But if you don't know what you want, it's a physical store with a physical salesperson still has some advantages. Obviously, lots of people do it on Amazon all the time. But you very often want to ask questions and help get help narrowing what you're going to look for. And as you keep asking new questions, having a whole bunch of different options presented to you. And I think AI and agentic commerce are going to change the experience online where that experience where you're narrowing what you want when you don't know is going to get better online than it even is in physical environments.

Now we obviously have our own efforts here in agentic commerce. We have Rufus, which I talked about in my opening comments, which is continuing to get better and better and used more broadly. And we have features like Buy for Me where we will surface on Amazon, even items that we don't stock that other merchants have. And then if customers want us to go and buy it for them on those merchants websites, we will do that. And both of those have been successful for us. But we're also having conversations with and expect over time to partner with third-party agents. And I think that it reminds me in some ways of the beginning of search engines many years ago being sources of discovery for commerce. And you had to kind of figure out the right way to work together.

And today, search engines are a very small part of our referral traffic and third-party agents are a very small subset of that. But I do think that we will find ways to partner. We have to find a way, though, that makes the customer experience good. Right now, I would say the customer experience is not -- there's no personalization. There's no shopping history. The delivery estimates are frequently wrong. The prices are often wrong. So we've got to find a way to make the customer experience better and have the right exchange value. But I do think that the exciting part of this and the promise is that AI and agentic commerce solutions are going to expand the amount of shopping that happens online. And I think that's really good for customers, and I think it's really good for Amazon because at the end of the day, you're going to buy from the outfit that allows you to have the broadest selection, great value and continues to deliver for you very quickly and reliably. And I think that bodes well for us.

And our final question comes from the line of Colin Sebastian with Baird.

I guess first on AWS, following up there. How much of this acceleration is driven by core infrastructure versus AI workload monetization? And I think part of it is trying to understand how important newer services like AgentCore are becoming and bringing enterprises to AWS to build agents? And then I guess, secondly, regarding the acceleration in advertising, if you could potentially disaggregate the core advertising contribution versus DSP and Prime video. That would be helpful as well.

I'll start on the AWS side, we are seeing -- we're really pleased with the results from this quarter, 20% year-over-year on a annualized run rate of $132 billion is unusual. And we have momentum. You can see it. And we see the growth in both our AI area, where we see it in inference. We see it in training. We see it in the use of our Trainium custom silicon. Bedrock continues to grow really quickly. SageMaker continues to grow quickly. And I think that the number of companies who are working on building agents is very significant. I do believe that a lot of the value that companies will realize over time and AI will come from agents.

And I think that building agents today is still harder than it should be. You need tools to make it easier, which is why we built strands, which is an open source capability that lets people build agents from any model that they can imagine. But even more so, when you talk to enterprises or companies that care a lot about security and scale. They're starting to build agents, and they don't really feel like they've got -- they've had building blocks that allow them to have the type of secure scalable agents that they need to bet their businesses and their customer experience and their data. And that's why -- that was really the inspiration behind AgentCore was to build another set of primitive building blocks like we built in the early days of AWS, where it was compute and storage and database. We defined a set of building blocks that you needed to be able to deploy agents securely and scalably that we provide in AgentCore. And then when we talk to our customers, it really resonates. There is not anything else like it, it's changing their time frame and their receptivity to building agents, and it's very compelling for them.

So I do think the combination of what we're doing to enable agents to be built and run securely and scalably as well as some of the agents that we're building ourselves that our customers are excited about are compelling for them. And I think the other place we see a lot of growth in AWS also is just the number of enterprises who are -- who have gotten back to moving from on-premises infrastructure to the cloud. And we continue to earn the lion's share of those transformations. And I look at the momentum we have right now, and I believe that we can continue to grow at a clip like this for a while.

I think on the advertising side, that is also an area where I think collectively, we feel very pleased about the progress. Every single one of our advertising offerings this quarter grew in a meaningful way. I think there's a few things going on for us. We have what I think of as a pretty unusual full funnel offering. And if you look at the top of the funnel, which typically tends to be awareness building and broad scale to be able to use our own Prime Video and our live sports capabilities as well as going all the way down to the bottom of the funnel at point of sale being able to use sponsored products, that's -- most people don't have a full funnel offering as robust as that. And then when you layer on top of it, the combination of the audience curation and development we can do, along with the advantage measurement, it just all leads to a return on advertising spend is very unusual.

And I think there are multiple places where we can expect to continue to grow. One is in our stores business. I still think if you look at the worldwide market segment share of retail, still 80% to 85% of it lives in physical stores. And that equation is going to flip over time. And I think AI is going to only accelerate that. So I think we have an opportunity -- a significant opportunity still in our existing stores. And then I think video, we've only been at this for a little bit of time, but it's already a very large amount of advertising revenue, and we're still relatively early stage. I think that will continue to be a big area of growth. And then as you referenced, the amount -- their demand-side platform or Amazon DSP, that is growing really quickly as well. And some of it had to do with the fact that we had some features. We always had a number of the core components people wanted around some of our properties, the measurement capabilities, Amazon Marketing Cloud, but we lack some features for a while as we were building out our DSP that customers told us mattered and the team over the last 20 months have closed those gaps in a very significant way so that now people feel like our DSP is fully featured.

And then you look at some of the partnerships that we've done, the Roku partnership gives us the largest connected TV footprint in the U.S. And you layer on top of that, what we've recently done in providing our DSP customers, the opportunity to integrate with the ad inventory in Netflix and Spotify and SiriusXM, it's powerful. And so we are growing very quickly on the demand side platform. So very optimistic about what we're doing there. We've continued work to do, obviously, but I don't think we're close to being able to grow there.

Thanks for joining us on the call today and for your questions. A replay will be available on our Investor Relations website for at least 3 months. We appreciate your interest in Amazon and look forward to speaking with you again next quarter.