Eugene Ng

Amazon $AMZN 3Q26 Earnings - Rev $201.9b +12% ↗️🟡 - GP $107.4b +17% ↗️🟢 margin 53.2% +238 bps ✅ - EBIT $28.1b +61% ↗️🟢 margin 13.9% +423 bps ✅ - Net Inc $20.7b -2% ↘️🔴 margin 10.2% -152 bps ↘️🔴 - OCF $42.4b +19% ↗️🟢 margin 21.0% +129 bps ✅ - FCF $-34.8b -- ⤴️🟢 margin -17.3% -1749 bps ↘️🔴 Revenue by Segment - Product sales $75.3b +2% ➡️🟠 - Service sales $126.6b +19% ↗️🟢 - Online Stores $68.5b +2% ➡️🟠 - Physical Stores $5.7b +2% ➡️🟠 - Third-Party Seller Services $46.5b +9% ↗️🟡 - Advertising $19.6b +11% ↗️🟡 - Subscription $14b +11% ↗️🟡 - AWS $46b +39% ↗️🟢 - Other $1.6b +16% ↗️🟢 - WW Paid Units +12% ↗️🟡 - WW Shipping Costs +11% ↗️🟡 North America - Rev $112.9b +6% ↗️🟡 - EBIT $8.7b +82.1% ⤴️🟢 margin 7.7% +322 bps ✅ International - Rev $43.1b +5% ↗️🟡 - EBIT $1.6b +31.8% ↗️🟢 margin 3.7% +73 bps ✅ AWS - Backlog $578b +189% YoY +17% QoQ 📶🟢 - ARR Rev $184b +39% ↗️🟢 - Rev $46b +39% ↗️🟢 - EBIT $17.8b +55.4% ↗️🟢 margin 38.7% +401 bps ✅ Mgmt Guide - 4Q26 Net Sales $246b +16% ↗️🟢 - 4Q26 Ad Rev +18% (high-teens) ↗️🟢 - 4Q26 EBIT $33.6b +34% ↗️🟢 - FY26 capex $220b unchanged 1 | “Lower” e-commerce sales due to “tough comps” with Prime Day occurring in Q2 this year instead of Q3. Q2 had “higher” and Q3 had “lower” growth. As a reminder, Prime Day timing shifted this year, with the event occurring in Q2 for most large countries. In 2025, Prime Day occurred entirely in Q3. Excluding this shift, third quarter year-over-year growth would have been approximately 400 basis points higher. 2 | Net income “declined” due to tough comps as there was $9.5bn of pre-tax gains from Anthropic in 3Q25. Net income decreased to $20.7 billion in the third quarter, or $1.89 per diluted share, compared with $21.2 billion, or $1.95 per diluted share, in third quarter 2025. This decrease is driven primarily by pre-tax gains of $9.5 billion in third quarter 2025 included in non-operating income (expense) from our investments in Anthropic. 3 | Q3 remains strong for Amazon, where AWS is just exploding. Q3 was another strong quarter for Amazon. In AWS, our plan is working. Despite lapping a comparable YoY period where we added a lot of capacity, AWS grew 39.3% YoY accelerating for the sixth straight quarter, our fastest growth in nearly 5 years. Backlog is $578 billion, growing triple digits YoY and well diversified. 4 | AWS is installing accelerators as fast as they can, and customers are just gobbling it up. Operating margins remain strong on the core side and rapidly improving on the AI side. 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. 5 | While Amazon is not guiding for AWS EBIT margins, the improving profitability is a result of disciplined efficiency gains, capacity optimization, and operating leverage. let me start with the operating margin outlook for AWS. So we're not guiding out 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 output, it's a result of a lot of disciplined efficiency gains, capacity optimization, closely managing our fixed cost. 6 | Even with higher memory costs and depreciation, AWS continues to accelerate their capex payback and has conviction that it will yield more revenue and free cash flows with strong ROICs. Even with heightened component costs like memory and 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. 7 | AWS has a compelling flywheel, much better than that of the neoclouds, which are just primarily hosting GPUs and renting short-term data center capacity. We have a compelling flywheel spinning in AWS. Borrowing from Jim Collins' flywheel definition Good to Great, where if you do A, you almost can't help but do B and if you do B, you almost can't help but do C and so on. In AWS, having the leading functionality, security 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, posttraining and tool usage, which can't help but drive reinvestment back into our broader AWS functionality. This is a different model than primarily hosting GPUs or renting short-term data center capacity. This is a long-term platform strategy that broadly helps customers invent and evolve their customer experiences, operate stably and build a reliable partnership they can grow for many years. 8 | 85% of global IT spend remains on-premises, and AWS is still riding a 10- to 20-year secular tailwind of which IT spend will shift to the cloud. 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 linchpin 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 learned to run it very efficiently over the last 20 years. 9 | The acquisition of DuckLabs will allow DuckDB to become the universal SQL dialect of AWS storage and analytics, allowing customers to have zero migration or portability cost between query engines. With our acquisition of DuckLabs, DuckDB will become the universal SQL dialect of AWS storage and 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. 10 | Bedrock is experiencing explosive growth, with spending 7x and token volume 17x YoY. 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 OpenAI models to Bedrock and usage is growing quickly. We also just added Bedrock managed versions of Kimi, GLM and SpaceX models. 11 | Customers keep choosing Bedrock because their data remains secure and private. Inputs and outputs are never shared with model providers or used to train their 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 so 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 at rest. 12 | AWS Continuum, which is used to discover, prioritize, validate, and remediate code vulnerabilities, and machine speed are already being used by numerous customers with interests unlike anything before. As frontier models get more powerful, they're surfacing 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 code's being written, Anthropic's Claude Code, OpenAI's Codex, and AWS's Kiro. The interest from customers exceeds anything we've ever seen for a new security service. 13 | Customers continue to use numerous AWS agent products, from Agent Core (purpose-built agents), Agent Connect (contact center agents), to Amazon Quick (AI work companion). Agents continue to be deployed at accelerating rates, many of which are being built on Bedrock AgentCore's building blocks. Tens of thousands of customers currently use AgentCore, deploying 2x more agents with it in Q3 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 that customers love Kiro for coding, Amazon Connect for contact center agents and Amazon Transform for migration. They're all iterating quickly and picking up customer momentum. As is Amazon Quick, 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. 14 | Amazon chip business grew triple digits. Revenue commitments for Graviton CPU and Trainium ASICs increase by more than 20x. Customers are saving more from using Amazon's silicon. Finally, our chips business continues to grow triple-digit percentages YoY and revenue commitments for Graviton and Trainium have each increased more than 20x YoY. AI continues to be a tailwind for Graviton because posttraining 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 15 | Despite the new generation of chips, demand for prior generations remains very strong, and continues to sell at a premium. Trainium now available outside of AWS. As we launch new generations of chips, demand for prior generations remained strong. After our launch of Graviton5, backlog for prior generation Graviton instances still more than doubled QoQ. And short-term capacity for prior generation Trainium instances sell at a premium compared to the first half of the year. With Trainium 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. 16 | AWS’s investments in custom chips and networking is building a cost-efficient infrastructure. Expect AWS margins to fluctuate QoQ. 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. 17 | Amazon’s own chips with Trainium and Graviton CPUs are highly price performant than GPUs and other x86 CPUs, and that is attracting other customers. Trainium, 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 posttraining and tool calling and we have the leading CPU chip in Graviton, which is about 30% to 40% more price performant than the other x86 providers. So it's -- those things are really driving very significant growth. 18 | Amazon is focusing on their own silicon like Trainium, to lower the inference costs. 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 inference down. That is a huge mission for AWS. It's one of the significant reasons why we're building our own custom silicon in Trainium is 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. 19 | Inference accounts for well over half of AWS AI revenue and continues to grow. 2026 cash capex remains unchanged at $220bn, the majority of which is for AWS. Expect this to grow significantly in 2027. AWS has a clear line of sight to attractive financial returns. Enterprises are still very early in using inference at 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. And 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. 20 | AWS can grow >40% in 2027 and 2028 and reach $1T revenue in the next 5-10 years (implied ~18-40% CAGR). AWS will be another trillion-dollar revenue business alongside the Amazon e-commerce business. 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 it's 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 ads that's growing at this pace. 21 | Amazon is on path to more than 2 businesses with >$1T of annual revenues. But the growth path of AWS will vary across quarters but be directionally much higher. 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…There has never been a trillion dollar revenue business, and we may have two of them in the next 5-10 years. 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. 22 | AWS’s capacity is reserved well in advance by customers, and they have good visibility into the next few years into 2027, 2028, and 2029, which is why they have confidence in AWS hitting $1T in revenues in the next 5-10 years. 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 1 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 on 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. 23 | AWS just open-sourced TorchNeuron so that developers can run existing PyTorch code without rewriting it. we just open-sourced TorchNeuron so developers can run existing PyTorch code, which most already use on Trainium without rewriting a thing. 24 | Think that AI models will not be a winner take most in token volumes. It will become increasingly commoditized where 6-8 models will be largely as good as other or 95%+ of the workloads. 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 been true in analytics. 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 at reasoning and some will be better at coding and some will be better at 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. 25 | Customers want different models, and that's why Amazon chose to have both closed and open models. 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 that 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 to 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's 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 models. And so I believe that will happen over time as well. 26 | Amazon has to continuously think about whether to use that compute internally for the long-term or sell it externally for higher near-term revenues. 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. 27 | Amazon expects a significant business selling their own chips externally (like Google)n from late 2027 onwards. 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. But 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 building a separate business and selling external chips and selling external chips really is a very different business than your cloud infrastructure business. 28 | Amazon continues to grow meaningfully faster than any retail competitor, increasing their market share. Year-to-date, units have grown nearly 15% YoY, 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. 29 | Rapidly improving ecommerce profitability by almost 1000bps has enabled Amazon to reinvest it to make the customer experience even better. 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. 30 | Amazon continues to focus on fast delivery and lowering their cost to serve. 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 Amazon Now around the world, is now available in over 450 towns and cities, and the number of items ordered through Amazon Now 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.

Alphabet $GOOGL 3Q26 Earnings - Rev $127.6b +25% ↗️🟢 - GP $76.2b +25% ↗️🟢 margin 59.7% +12 bps ✅ - EBIT $40.6b +30% ↗️🟢 margin 31.8% +129 bps ✅ - Net Inc $26.5b -24% ↘️🔴 margin 20.7% -1343 bps ↘️🔴 - OCF $42.3b -13% ↘️🔴 margin 33.1% -1416 bps ↘️🔴 - Capex $51.4b +114% ↗️🟢 - FCF -$9.1b ↘️🔴 margin -7.1% -3101 bps ↘️🔴 Revenue by Segment - Google Services $96.3b +11% ↗️🟡 - Google Advertising $82b +11% ↗️🟡 - Google Search $63.4b +12% ↗️🟡 - YouTube Ads $11.2b +9% ↗️🟡 - Google Network $7.5b +1% ➡️🟠 - Google Subscriptions $14.3b +11% ↗️🟡 - Google Cloud $30.7b +103% ⤴️🟢 - Other Bets $0.4b +8% ↗️🟡 Google Services - Rev $96.3b +11% ↗️🟡 - EBIT $39.2b +16.8% ↗️🟢 margin 40.7% +215 bps ✅ Google Cloud - Rev $30.7b +103% ⤴️🟢 - EBIT $11.1b +209.4% ⤴️🟢 margin 36.2% +1250 bps ✅ - Backlog $561bn +256% ↗️🚀 Other Bets - Rev $372m +8% ↗️🟡 - EBIT -$2b ↘️🔴 margin -550.3% -13573 bps ↘️🔴 Revenue by Geography - US $67.7b +39% ↗️🟢 - EMEA $32.7b +9% ↗️🟡 - APAC $19.9b +12% ↗️🟡 - Other Americas $7.1b +17% ↗️🟢 1 | Revenues grew 25% and are now at a $500 billion ARR. Continues to have strong momentum, and AI investments are increasing usage and overall query growth. Alphabet revenues grew 25% YoY. We are now at $0.5 trillion annual run rate. That's remarkable growth. Search has strong momentum. Our AI investments are driving increased usage and overall query growth, including commercial queries. In fact, our YoY query growth rate is the highest it's been in the last 5 years. 2 | Core Google revenues continued to grow fairly strongly, across advertising, search, YouTube, and subscriptions. Google Services revenues increased 11% to $96.3 billion, reflecting strong growth in Search and subscriptions. Total advertising revenues were up 11% with growth across all major verticals….Google Search and other advertising revenues increased by 12% to $63.3 billion, with retail driving the largest contribution followed by finance. YouTube advertising revenues increased 9% to $11.2 billion, with brand driving the largest contribution and notable strength in living room. Network advertising revenues of $7.5 billion were up 1%. Subscription, platforms and devices revenues increased 11% this quarter to $14.3 billion, driven by strong growth in YouTube subscriptions, particularly in YouTube Music and Premium and strong demand for AI plans in Google One. 3 | Noise in slower growth of search/advertising revenues from FX, APAC advertisers Impacted by the EU eliminating its diminutives exemption and ad spend normalizing after a strong Q2 driven by the FIFA World Cup, and lower monetization of AI mode. Search grew 12% YoY to $63 billion. And when you look at the change in growth rate from Q2 to Q3, there were a few drivers here. So first, there was FX. It was the single largest driver of Search deceleration relative to Q2. Second, there were vertical-specific dynamics. In retail, some of our APAC advertisers were impacted by the EU eliminating its de minimis exemption. In media and entertainment, spend normalized after an unseasonably strong Q2, driven by the FIFA World Cup, which had more matches and therefore, more advertising budgets allocated in June than actually in July. It's also -- to the next part of your question, it's worth noting that the strong growth in AI Mode queries had an impact because of the lower monetization we were seeing in AI Mode. To be clear, we feel very good about our trajectory here. Our ad experiments have shown promising results, and we're now scaling ads within both AI Overviews and AI Mode to more users and more queries, including, as I mentioned, international markets and multiturn ads in the U.S. 4 | Higher costs due to TPU sales increase in depreciation, content acquisition cost, and technical infrastructure operation cost, and an $8.2bn unrealized loss from SpaceX. Other cost of revenues was $35.1 billion, up 32%, driven by hardware costs primarily related to the sales of TPU systems, increase in depreciation, content acquisition costs and technical infrastructure operations costs. Total operating expenses were up 20% to $35.6 billion. Other income and expenses was negative $8.2 billion due primarily to unrealized loss from our investment in SpaceX from a decline in its stock price in the third quarter. 5 | Saw lower OCF due to the buildup of TPU inventory and the $5.2 billion cash fine. We generated strong operating cash flow of $42.3 billion in the third quarter and $179.6 billion for the trailing 12 months. Operating cash flow was negatively impacted by the buildup of inventory for TPU system sales as well as a $5.2 billion cash payment for a 2018 EC fine following the conclusion of the appeal process. 6 | 2/3 of capex was for servers and networking equipment, and the remaining was for data centers. CapEx was $51.4 billion in the third quarter with the vast majority of the spend in technical infrastructure to support our investments in AI. Approximately 65% of our investments in technical infrastructure in Q3 was related to servers and networking equipment, which generally have useful lives of 6 years. The remaining 35% was for data centers, which have useful lives of 7 to 40 years. We had negative free cash flow of $9.1 billion in the third quarter, driven by our investments in CapEx. Free cash flow in the third quarter was also affected by the aforementioned inventory build to support TPU system sales and payment of the EC fine. 7 | Google has maximum contractual obligations of $1,076bn, $924bn of future purchase commitments and $152b of other leases and obligations, most of which have some variable pricing and volume flexibility, which could result in a lower spend. We have also entered into multiyear agreements to secure capacity for technical infrastructure and inventory as well as leases and energy. At the end of the third quarter, we had future purchase commitments of $924 billion and pre-commencement leases and other obligations of $152 billion. These dollar amounts reflect the maximum contractual obligations. However, a number of these agreements have variable pricing and volume flexibility, which could result in lower spend. 8 | GCP revenues grew 103%. Seeing strong customer demand for Google TPUs. Google Cloud revenues grew 103%. Our AI infrastructure and solutions are the major drivers for Cloud growth and our customers see us as the leading provider here, so there's huge demand. We are working hard to bring more capacity online to meet this long-term demand. At the same time, we are seeing strong demand for our TPU systems across a wide range of customers, and we expect that to continue as we scale 9 | TPU system sales were a meaningful contributor, expected to scale in Q4, followed by a meaningful revenue ramp in 2027. TPU system sales were also a meaningful contributor. GCP accounted for the vast majority of the revenue generated in Q3. The cloud revenue growth rate, excluding the impact of TPU systems, was again incredibly strong. In Google Cloud, we expect TPU system sales to continue to scale in Q4, followed by a meaningful revenue ramp in 2027. 10 | Third-party bridging capacity only came online at the end of Q3 and had limited negative impact on margins. Expected negative impact to be larger from Q4 onwards, and more near-term as they pay up to get long-term customers. The third-party bridging capacity, which we discussed on our prior call, came online only at the very end of Q3 and therefore, had limited negative impact to our margin in the quarter. With respect to Cloud margins, given the supply constrained environment, we started utilizing third-party compute capacity at the end of Q3 and will ramp usage in Q4. We expect margin headwinds in the near term as we utilize this capacity. 11 | Google Cloud backlog is $561b up from $514b, expecting higher capex to match the growth in backlog. Google Cloud's backlog was $561 billion in the third quarter, with the majority of the backlog related to typical GCP contracts. Moving to investments. Our full year 2026 CapEx guidance range of $195 billion to $205 billion remain unchanged. We continue to expect our CapEx to increase significantly in 2027, and we'll provide further details on our Q4 call. In terms of expenses, the significant increase in our investment in technical infrastructure will continue to weigh on the P&L in the form of higher depreciation expenses and related data centers operation costs such as energy. 12 | Google expects to generate an attractive ROIC that is meaningfully higher than the cost of capital, which will result in growing OCF. let me say a few words on how we think about return on invested capital and free cash flow. We're in a unique position to be able to monetize our AI investments across many products that reach billions of users and enterprises. Our AI investments in Search are driving our highest query growth in half a decade. Google Cloud has generated tens of billions of dollars in AI revenues year-to-date, and Google One is our fastest-growing subscription products, driven by demand for AI plans. Throughout this investment cycle, we expect to generate an attractive ROIC that remains meaningfully above our cost of capital and anticipate generating strong cash flow from operations 13 | Customers are using Google's AI models more than ever. 1P APIs are processing 32B tokens per minute, +40% QoQ. Google is back at the frontier. On our AI models, customers are using our models more than ever. Our first-party model APIs are now processing nearly 32 billion tokens per minute, 40% more than last quarter and more than 4x what we processed a year ago. And with Gemini 4 Argon and now Carbon, which we announced today, we are excited to be leading at the frontier. The Gemini app is now at 1 billion users and is the fastest-growing product in Google's history. 14 | Saw strong demand for Argon, ramped 3.5x faster than Gemini 3. Using Argon internally and shipping 35% more code with 30% fewer errors. Starting with AI, highlighted by our Gemini 4 series of frontier models. We are seeing strong demand and usage for Argon, which we launched 2 weeks ago. In fact, on our developer platforms, usage in the first week ramped up 3.5x faster than Gemini 3, even with capacity constraints. Argon is proving to do really well in areas like finance, legal, cybersecurity and science. Developers and cloud customers are already using it for things like untangling legacy code bases, accelerating scientific research, robotics, discovering product improvements and domain-specific tasks requiring in-depth reasoning. And at Google, just in the last few weeks, teams across Chrome, Ads and Search have used Argon to write and optimize code much faster, shipping 35% more code with 30% fewer errors. Argon is just the start of the Gemini 4 series. 15 | Carbon brings Google right back to the frontier for AI models. Today, we announced Carbon. It's our strongest frontier model with state-of-the-art cybersecurity and long horizon capabilities. It ranks #1 on the recently updated artificial analysis intelligence Index at 59. Carbon outperforms Argon across almost every domain like full stack web development, multimodal understanding, complex knowledge work and scientific reasoning. We are rolling Carbon out first to a subset of trusted cyber defenders before we expand access, and it's being tested by the U.S. government. 16 | Genie 4 which simulates safe, realistic, interactive worlds with native audio in real time will be an important building block for training robots and agents. Beyond our Gemini 4 series, yesterday, we announced Genie 4, our most advanced world model yet that simulates safe, realistic, interactive worlds with native audio in real time. This is an important building block for training robots and agents. 17 | Recursive self-improvement (RSI) is helping Google to increase their training and launch velocity for AI models. This rapid progress underscores our incredible model momentum. We are at the point where AI is helping us increase our training and launch velocity. I would expect the next 12 months to see continued acceleration because of RSI. We have more models coming soon, including more efficient versions to help developers optimize for speed and cost and to deploy across our billions of users. 18 | Google Cloud now has 18 product lines with more than $1 billion in annual revenue. TPU sales were a meaningful contributor. Google Cloud now has 18 product lines with more than $1 billion each in annual revenue. And our growth this quarter was driven by strong growth in Google Cloud Platform with strength in both AI and core GCP. TPU system sales were also a meaningful contributor. 19 | Google worked closely with Meta to deliver TorchTPU, allowing developers to port workloads easily from GPUs to TPUs. To make it easy for developers to port workloads to TPUs, we recently worked closely with Meta to deliver TorchTPU, which allows users to train, serve and deploy PyTorch models on TPUs, using the tools they're already familiar with and rely on. Nearly 100 customers use TPUs today, and we're seeing strong and growing demand from many others. 20 | Google is seeing new customers, like Apple, capital market companies, universities, and public sector companies use TPUs. GCP recently entered into partnerships to deliver TPUs to Crux AI and Oracle. In addition to our work with Anthropic, both directly and via Broadcom and public reference customers like Apple, who are scaling pretraining of their foundation models on the latest generation of TPUs, we're also working with new frontier model companies, including Mirendil and Physical Intelligence; capital markets firms, including Citadel Securities and Jump Trading; universities like Caltech; and public sector service providers, including Cassava and Cirrascale. Given continued strong and accelerating demand, we've recently entered into partnerships to deliver TPU systems with both Crux AI and Oracle, with more in the works. 21 | Google now has over 100 significant TPU customers across digital natives, AI labs, traditional enterprises, and financial and capital markets customers. on the demand in Cloud, we are always managing our portfolio to ensure we have a breadth of customers, and we have different types of customers from digital natives, AI labs, traditional enterprises, financial and capital markets customers. So it's a broad set of customers. Many of them have either started with using TPU systems, for example, in our cloud. We have over 100 significant TPU customers who start in our cloud. They learn to use TPUs. And remember, a TPU system is a combination of hardware, system software, compilers, the whole thing assembled as a system, including optical networking and other things. We've taken a very deliberate approach on how we're bringing TPUs to customers. The first one is to build the right silicon. So we have 8t for training, 8i for inference. They both offer substantial price performance and performance improvements relative to prior generations. 22 | Google has now optimized the software, PyTorch in TPU, and are broadening distribution for customers to use TPUs, not just in Google Cloud but also in their own data centers and to other providers like Oracle and Crux AI. Second, we've optimized the software. The optimization in the software is not just to make the systems run better and to give you great performance but also to see -- make portability seamless. Recently, we introduced, for example, PyTorch, which opens up a much larger market for us by optimizing PyTorch in the TPU. Third, we are broadening distribution. So in the past, you could use TPUs in our cloud. Now we make for customers who want TPU systems in their data centers, we're making TPU systems available in their data centers. We're also broadening availability to other providers. For example, we announced a partnership with Oracle and we've launched along with Blackstone an agreement for Crux AI, a joint venture to offer customers TPU systems. 23 | Currently not breaking out any margins for TPU systems, focused on improving margins as they scale the TPU business. On TPU systems and margins, we don't break out margins for individual products or infrastructure components. However, we see several benefits from designing and manufacturing TPU systems. So first of all, to remind you, the TPU system is not just a chip, but it's a complete system. Given the price performance benefits of TPU systems relative to competitors, growing demand for accelerated compute and the scarcity of accelerated compute, we're very focused on continuing to improve margins as we scale the business. 24 | TPUs are manufactured only after contracts are signed with customers. All inventory is associated with committed contracts. By selling TPU systems, it offsets the need for Google to pre-build data centers to host these TPU systems and allows them to scale faster. Two additional points. First of all, when Anat mentioned that we procure components to build inventory, we only do that after we have signed contracts. So all of the inventory that -- for TPU systems are associated with committed contracts. Additionally, by putting TPU systems in third-party facilities and in customer data centers, it allows us, one, to broaden our addressable market. Some -- for some industries, they have data in their premise that they cannot move to a cloud, and it allows us to bring accelerated compute to where the data sits. Additionally, it also offsets the need for us to have to prebuild facilities to host these TPU systems. 25 | Gemini is a powerful accelerator for Google Cloud. Gemini is also driving acceleration for our database, analytics and cybersecurity products. Gemini can now help users find, analyze and visualize information from structured and unstructured data. Gemini uses our knowledge catalog to find the right sources and definitions of information, improving accuracy, lowering token cost and making responses more reliable. Gemini uses our Lakehouse to access information from multiple systems with 0 copy. To respond to agentic analysis, BigQuery improved query speed over 40% YoY while reducing query processing cost by over 50% YoY. 26 | More than 2,600 customers have used at least $100 billion in tokens, and nearly 90% of Fortune 100 use Gemini Enterprise. Over the past 12 months, more than 2,600 customers have used at least 100 billion tokens and nearly 75 at least 10 trillion tokens on Gemini Enterprise. Nearly 90% of the Fortune 100 use Gemini Enterprise, and customers who use Gemini Enterprise in Q3 2025 more than quadrupled their spend with us year-over-year. One of the important ways in which customers are using Gemini is for knowledge work. Across all our AI services, we have sold more than 100 million seats for knowledge work. Gemini Enterprise now includes the Gemini agent that is deeply integrated across our entire cloud portfolio. 27 | People using Google's AI features have 3x longer queries than traditional searches, and 1 in 5 use voice, images, or documents. Let's turn to Search, which is bringing the benefits of AI to more people than any other product in the world. Just like our shift to mobile over a decade ago, we are focused on building the best Search user experience, followed by monetization. We are building one seamless experience across AI Overviews, AI Mode and traditional search. And people are using our AI features more searching in entirely new ways. Their queries are longer, personalized and multimodal. Queries in AI Mode are on average 3x longer than traditional searches, and 1 in 5 AI Mode searches globally use voice, images or documents. We have now crossed over 50 billion visual searches per month. One area I'm excited about is helping people get things done, be it booking reservations, filling in forms, scheduling appointments or making purchases. 28 | Google wants to bring agents as a unifying layer across all their products. People are already using the Gemini app to do more than ask questions. They're using it to do things, including complex multistep agentic queries. Our Gemini 4 series has been a really big step forward here and has already been handling these complex tasks really well. Agents will operate as a unifying layer across our products, so you will see us extend these agentic experiences beyond Search in the Gemini app to Chrome, Android and our Workspace products as well. 29 | Google first wants to roll out Gemini agents for enterprise, then later on to consumers. We have been hard at work, and we decided to roll it out to enterprise first, which we announced with our Gemini agent for enterprise, essentially because they have the highest bar for security, governance and controls. And it's been well received, and we are in the process of expanding that. Later this year, we will bring that same experience to consumers, obviously, starting with the 1 billion users in our Gemini app. We already see people in the app moving beyond chat to what are tasks and agent-like experiences. 30 | YouTube continues to lead as the number one streaming platform in the U.S. in terms of watch time. The living room accounted for one-third of total YouTube ad revenues. On to YouTube, which continues to lead in the living room as the #1 streaming platform in the U.S. in terms of watch time. It's also seeing good subscriber growth as I mentioned. At Made on YouTube last month, we announced more than 30 products and features for creators and viewers. One great example is custom feeds where you can simply pick a topic you love and get a tailored video home feed based on that topic…The third part is monetization starting with ads. Our focus remains on delivering value across the entire funnel. In Q3, living room accounted for over 1/3 of total YouTube Ads revenues. 31 | Amazon is joining YouTube shopping. Creators can now tag Amazon products directly on YouTube. More than 1.3 million channels are enrolled in YouTube Shopping globally. Building on this momentum, we recently announced that Amazon is joining YouTube Shopping. U.S. creators can now tag Amazon products directly in their videos, shorts and live streams, driving direct conversions. Importantly, our subscription ecosystem continues to complement our advertising business. 32 | Integration of Gemini models into Google's ad infrastructure is improving how they decode user intent and driving growing ad relevance. First, ads quality. The integration of Gemini models into our ads infrastructure is improving how we decode user intent and drive ads relevance. For example, in Q3, we made the biggest update to product listing ads in a decade, adding new AI capabilities to help users make more confident purchase decisions. Already, we've seen satisfaction increase by 19% and dissatisfaction dropped more than half. We're also seeing AI models enable better prediction of clicks and conversions, stronger overall performance and deeper personalization of ads. For instance, through several innovations and personalization modeling, we delivered 18% incremental advertiser value and an 8% increase in user sentiment on Discover ads in 2026. Overall, AI is accelerating our ads development and drives the majority of our product launches. 33 | AI is also accelerating advertiser tools, allowing every advertiser, regardless of size, to seamlessly participate in all of Google's AI surfaces. Second, advertiser tools. We're ensuring every advertiser, from global brands to small businesses, can seamlessly participate in our AI surfaces. AI Max and Performance Max continued to see strong adoption. To give marketers even more control, we're now introducing a way to test different budgets and ROI targets across multiple search campaigns in a single AI Max A/B test. To drive better performance, we integrated our AI creative tools directly into AI Max. In Q3 alone, Gemini models helped advertisers generate over 1.5 billion campaign assets, up over 7x YoY.