Prepared remarks
Welcome, everyone. Thank you for standing by for the Alphabet Second Quarter 2026 Earnings Conference Call. I would now like to hand the conference over to your speaker today, Jim Friedland, Head of Investor Relations. Please go ahead.
Thank you. Good afternoon, everyone, and welcome to Alphabet's Second Quarter 2026 Earnings Conference Call. With us today are Sundar Pichai, Philipp Schindler, and Anat Ashkenazi. Now I'll quickly cover the safe harbor. Some of the statements that we make today regarding our business, operations and financial performance may be considered forward-looking. Such statements are based on current expectations and assumptions that are subject to a number of risks and uncertainties. Actual results could differ materially. Please refer to our Forms 10-K and 10-Q, including the risk factors. We undertake no obligation to update any forward-looking statement. During this call, we will present both GAAP and non-GAAP financial measures. A reconciliation of non-GAAP to GAAP measures is included in today's earnings press release, which is distributed and available to the public through our Investor Relations website located at abc.xyz/investor. Our comments will be on year-over-year comparisons unless we state otherwise. And now I'll turn the call over to Sundar.
Hi, everyone. Thanks for joining us. We have exciting momentum. Alphabet revenue grew 24% year-over-year. Our AI investments are redefining what's possible across every part of our business. Our momentum starts with Search where people are adopting one seamless search experience across AI Overviews and AI mode. We saw 17% revenue growth in Search and Other and YouTube ads grew 13%. Cloud revenue grew 82% powered by strong demand for AI infrastructure and AI solutions and Cloud backlog grew to $514 billion. It's great to see the wide adoption of Gemini Enterprise with nearly 90% of the Fortune 100 using it. Today, I'll expand on our AI progress, our global product footprint, led by Search and YouTube, Google Cloud's leadership and our long-term bets. First, AI. Yesterday, we announced new models, Gemini 3.6 Flash and 3.5 Flashlight, which are cost-effective and highly efficient. We are seeing tons of demand for our workhorse Gemini Flash series because it hits the sweet spot of performance and cost. We also launched Gemini 3.5 Flash Cyber, which I'm really excited about; paired with our core vendor agent, it finds and fixes vulnerabilities and delivers performance at the frontier comparable to far bigger cyber models. GEMINI 3.5 Pro is currently in testing and our team is already building the next generation of models. We have started our most ambitious pretraining run yet for Gemini 4 and are excited by the progress we are seeing at the frontier. Demand for our models is translating to strong token usage across developers and enterprise customers, and we continue to be supply constrained, a sign of momentum and rapid adoption. More than 9 million developers are building each month with our models across our APIs and key developer products. Our model APIs are now processing approximately 22 billion tokens per month. That's up from 16 billion just a quarter ago. Additionally, this quarter, we launched Omni. It allows users to create anything from any input starting with video. Since launching at I/O in May, there's been a 40% increase in daily active users creating videos on the Gemini app. Our Gemini family of open models, small enough to run on local devices, are hugely popular. These models have been downloaded over 900 million times, and our latest GM4 models have been downloaded over 300 million times since launching in April. Our agentic development platform, Antigravity, allows anyone to build in the agent-first era. It is more than 2.4 million weekly active users. Antigravity is a powerful tool for users and enterprises and it's completely accelerated how we build internally. As just one example, a team in Chrome is now on track to accelerate delivery by 8x, compressing a two-year timeline into three months through model-driven refactoring. Moving on to our global product footprint, which brings AI to more people than any other company. Starting with Search. AI continues to drive an expansionary moment with new experiences resonating with users and driving growth in queries. As a big football fan, I was particularly excited to see Search usage hit an all-time high during the World Cup this year. This really highlights how much people turn to Google in moments that matter. We continue to make Search more helpful and intuitive. We recently brought together AI Overview and AI mode into one seamless search experience that combines our frontier capabilities with the best of the web. And we are continuing to incorporate more frontier capabilities into Search with agents, personal intelligence and notebooks. Our AI-powered features are driving increased search usage: since expanding AI mode globally last October, we have surpassed 1 billion monthly active users. And just like AI Overviews, AI mode is driving an incremental increase in search queries overall, and we are now sending billions of clicks to websites every week through AI features in Search. As we serve more of these queries, we have continued to drive efficiencies. Thanks to our engineering and hardware optimizations this quarter, we reduced the cost of AI mode responses to its lowest level since launch even as we have brought more advanced AI capabilities. Philipp will also talk more about how we are innovating the ads experience. Next, the Gemini app, which now has 950 million monthly active users with daily active users tripling in the last year. Users love new agent features like Daily Brief and our personalized agent, Gemini Spark, which is now available in the U.S. and internationally. We have been shipping helpful new features like this at an incredible clip. On to YouTube. Month after month, when a big event happens in the world, the world turns to YouTube. It's been incredible to see that over 1.7 billion unique viewers globally watched World Cup-related videos on YouTube during the FIFA World Cup 2026. We are also bringing the power of conversational AI directly into the YouTube experience. Ask YouTube uses our Gemini models to let people ask complex questions about individual videos, get quick takeaways and jump straight to moments in those videos. The early engagement is encouraging. More than 140 million users engaged with Ask YouTube on the watch page in June 2026. And we are bringing that Ask experience to the wider Search experience on YouTube. Next, Cloud. Our continued momentum is driven by our integrated AI portfolio, consisting of chips, models, data, security and agent platforms, all designed to work together. Gemini continues to be a key driver of growth and is deeply integrated across all of our cloud products, including Gemini Enterprise, data analytics, cybersecurity and Google Workspace. We are seeing strong diversified demand across products, customers, geographies and industries. Our product differentiation is driving expansion in three ways. We are winning new customers, more than doubling our acquisition velocity year-over-year. We are deepening our relationships with existing customers who are expanding their usage and exceeding their commitments by more than 50% — an acceleration over last quarter. And we are driving growth with partners with transactions on Google Cloud Marketplace growing over 7x year-over-year. One of the strongest parts of our growth comes from the rapid adoption of our Gemini Enterprise platform. It's differentiated with easy-to-use tools to build agents and automate processes, connectivity to enterprise systems, cost management and governance tools. In Q2, Agent Development Kit, our framework for building and deploying enterprise AI agents, reached nearly 70 million total downloads. As I said, nearly 90% of the Fortune 100 are using Gemini Enterprise. We have customers like PepsiCo for AI and analytics solutions, Intel to streamline core processes, HSBC for wealth management, Bell Canada for customer engagement, Macy's for commerce experiences, and Signal Iduna for knowledge management. More broadly, Gemini is transforming how millions of businesses use AI to build custom agents, automate processes, improve cybersecurity, manage customer relationships, streamline data analytics, collaborate effectively and more. All of this momentum is driving growth in our paid token usage. Nearly 500 cloud customers have each processed more than 1 trillion tokens in the last year and usage is so much deeper than that. Over the last 12 months, more than 2,000 enterprises consumed over 100 billion tokens. We are also seeing strong interest in our AI-powered security platform, which is differentiated because it integrates threat intelligence, cyber response prioritization with best-in-class automated scanning, code remediation and monitoring. Today, 90% of the Fortune 100 are Google Cloud security users. Nearly 90% of these customers are using AI-powered security features. And we have seen a more than 45% quarter-over-quarter increase in the number of AI workloads scanned and protected by our security platform. Our security tools are being used to protect critical infrastructure including financial services organizations such as Morgan Stanley, telecommunication providers such as TELUS, health care organizations such as Texas Children's, software companies such as Atlassian and several government agencies. And with our new Google AI Threat Defense, we are really excited to bring our new cyber model and tooling to help our customers defend against AI threats. All of this is powered by our leading AI infrastructure. We offer the industry's broadest range of accelerators from Google and NVIDIA, including the latest NVIDIA platforms and our TPUs, which deliver strong price performance. On top of this, let me mention three unique offerings. First, our Virgo network, which is designed to meet the needs of modern large-scale AI workloads. It allows customers to connect 1 million AI accelerators across multiple data center sites into a unified supercomputer. Second, our software stack has native support for JAX, PyTorch and TensorFlow, enabling workload portability across GPUs and TPUs. And third, our new agent-optimized Axion CPU provides better performance per dollar compared to peer offerings. We are seeing strong growth in demand for our AI infrastructure offerings from leading labs such as Inflection, next-generation AI builders including Kakao, financial services like Deutsche Börse Group, pharmaceutical companies such as Pfizer and Roche and robotics and spatial intelligence companies such as World Labs. We are continuing our strategy of investing in promising companies and internal efforts that bring pioneering AI research into the real world. We have seen exciting progress from our other bets. Waymo introduced its newest vehicle for public riders. This is the first vehicle powered by the sixth-generation Waymo Driver and we'll welcome more riders in the coming months. Wing has safely completed more than 1 million home deliveries, and we continue to grow our presence through partnerships with Walmart, DoorDash and Papa John's. In health and drug discovery, Isomorphic Labs raised over $2 billion to power its AI drug design engine, scale globally and advance its drug candidate pipeline. It's truly been an extraordinary first half of 2026 with much more to come. Thanks to all of our employees and partners worldwide for their incredible work this quarter. With that, Philipp, over to you.
Thanks, Sundar, and hello, everyone. I'll start with a review of Google Services, highlighting our momentum across Search, YouTube and partnerships. Google Services revenues were $95 billion for the quarter, up 15% year-on-year, primarily driven by Search. Search and Other delivered 17% growth with retail and finance driving the largest contributions. YouTube advertising revenues grew 13%, driven by direct response and brand. Network advertising revenues were down 1% year-on-year. Strong performance in Search and YouTube underscores how our investments in AI translate into measurable value for our users and advertisers. Starting with Search and Other. Revenues reached over $63 billion for the quarter. Sundar highlighted the momentum we see in Search, which directly impacts our advertising business. We continue to accelerate the deployment of Gemini across our entire ads infrastructure to boost performance in three areas I mentioned before: ads quality, advertiser tools and AI user experiences. First, ads quality. At Google Marketing Live, we showcased how Gemini improves query understanding, allowing us to find relevant ads for longer searches previously difficult to monetize. The core engine of our search ads relies on a dual prediction, delivering immediate utility for the user while maximizing measurable value for the advertiser. Gemini completely supercharges this capability. We use Gemini's advanced reasoning to decode the nuances of longer, more detailed queries. With shopping ads, for instance, we drove a 20% improvement in showing highly relevant ads, helping shoppers immediately find the best match. Second, advertiser tools. Take AI Max. It has become the core building block for advertisers to fully participate in our new AI experiences. It's out of beta and 0.5 million advertisers have already adopted it. Those who adopt our AI-powered campaigns like AI Max or PMax see an average of 15% more conversions or value on Search at a similar ROAS. AAA Auto Club Enterprises used AI Max to personalize creative assets and capture growth from increasingly detailed insurance searches. This led to a 17% improvement in conversion volume and an 11% decrease in cost per lead. We see strong adoption of our generative AI creative tools, which makes creative development easier, especially for SMBs. In fact, over half of our SMB customers globally use AI to create or optimize their creative. Third, new AI user experiences. At I/O, we showed how Search, including AI Overviews and AI mode, works as one connected experience. As you saw from our results, our performance across the holistic system is robust. We continue to be encouraged with monetization performance on queries that show AI Overviews, even as we've expanded overviews to more commercial queries. Across AI Overviews and AI mode, people are asking more specific and detailed questions, providing opportunities for more relevant ads. Within AI mode, we continue to test and deploy a range of new ad formats. For text ads, we improved performance by adding contextual cycling based on the conversation. Direct offers is gaining momentum with partners like IHG Hotels & Resorts, soon servicing special offers during trip planning. Highlighted answers and clearly marked sponsored links inside list responses are showing early user traction. Looking at commerce. In collaboration with the retail industry, we established the open-source Universal Commerce Protocol, UCP, as a new standard for agentic commerce. Merchants are rapidly adopting UCP with Target and Steve Madden now live, while new members have joined the UCP shopping and food tech councils to help steer its vision. We also announced Universal Card, allowing shoppers to add items from different retailers across Google services and buy in a single checkout. Now on to YouTube. As a FIFA preferred platform partner, we help fans around the world tune into the games. As Sundar mentioned, over 1.7 billion unique viewers watched World Cup-related videos and over 550 million watched them on their televisions. This made the FIFA World Cup 2026 the most viewed World Cup in YouTube history. Advertisers connected with fans through FIFA channel takeovers, Gemini-powered soccer theme sponsorships and game day mass heads and, of course, a huge congratulations to Spain for taking home the trophy. More broadly, on YouTube, we introduced an exclusive slate of creator shows making it easier for brands to tap into the fandom of creators. We launched custom sponsorships to put brands at the heart of the world's biggest moments as they unfold on YouTube using AI to dynamically surface videos tailored to a brand's desired moment. Brands continue to partner with YouTube creators to engage new audiences. Kate Spade reached Gen Z through a first-of-its-kind YouTube creator campaign and partnered with creators Li Tuan and Hanami by leveraging multiple format videos on YouTube; the brand drove a 3.25% brand lift and purchase intent. This unique convergence of brand equity and commercial action makes YouTube a powerful full-funnel platform. SMBs continue to fuel our direct response growth through campaigns like Demand Gen; they leverage our AI to scale visual storytelling across YouTube, Shorts and now Google Maps. To attract high-value shoppers across North America and Europe, outdoor brand Arterix used Demand Gen to achieve a 70% better return on their ad spend compared to other paid channels. Looking at monetization across YouTube, we're driving sustained growth across our key priorities. In the living room, we see continued momentum across both brand and direct response. With the launch of Buy with Google Pay, viewers can complete purchases directly on their CTV, turning the TV screen into a stronger performance surface. Shorts continues to deliver high-performing opportunities for social and video buyers alike. Beyond advertising, YouTube's subscription business, which thrives across living room screens, is growing faster than ads, particularly driven by YouTube Music and Premium. Internally, we leverage Gemini to transform how we work and how we serve our customers. 83% of our sales team uses Gemini-assisted tools weekly driving up to a 20% higher win rate when using customized pitch narratives. Our ads customer support teams use Gemini-powered agentic solutions that now autonomously address 75% of support cases, freeing them to solve our customers' most complex challenges. Similarly, our new agentic solutions for SMBs have expanded our reach by hundreds of thousands of new customers year-to-date. As always, ending with partnerships. We're seeing more businesses benefit from the unified strength of Google's AI stack. Booking Holdings expanded a multiyear cloud commitment and are partnering closely to advance our AI ad formats. It's also deploying Google's AI technology to enable new customer experiences like agentic dining reservations on OpenTable, helping restaurants get discovered and booked right when it matters most. In closing, I'd like to thank Google employees everywhere for their contributions to our success and our customers and partners for their continued trust. Anat, over to you.
Thank you, Philipp. My comments will focus on year-over-year comparisons for the second quarter unless I state otherwise. I will start with results at the Alphabet level and will then cover our segment results. I'll end with some commentary on our outlook for the third quarter and full year 2026. We had an outstanding second quarter, delivering our 12th consecutive quarter of double-digit revenue growth. Consolidated revenues were $119.8 billion, up 24% (or 23% in constant currency). Total cost of revenues was $45.9 billion, up 18%. Technology infrastructure was $16.2 billion, up 10%. Other cost of revenues was $29.8 billion, up 22%, driven by increases in depreciation, inventory costs, primarily from the sales of TPU systems to customers, and content acquisition costs largely for YouTube. Total operating expenses were up 27% to $33.1 billion. R&D expenses increased by 32%, driven by compensation from investments in AI talent as well as depreciation. Sales and Marketing expenses were up 18% due to investments to support the Gemini app in Search and compensation. And G&A expenses increased 24%, primarily driven by compensation and charges for certain legal and other matters. Operating income increased 30% to $40.8 billion and operating margin was 34%. Other income and expenses was $9.8 billion representing a substantial increase from the prior year, primarily due to unrealized gains in our equity securities portfolio. Net income and earnings per share increased significantly primarily due to the unrealized gains in Other Income & Expense I just mentioned. We generated strong operating cash flow of $39.1 billion in the second quarter and $185.7 billion for the trailing 12 months. CapEx was $44.9 billion in the second quarter, with the vast majority of the spend in technical infrastructure to support our investments in AI. Approximately 60% of our investment in technical infrastructure this quarter was in servers and 40% was in data centers and networking equipment. We had negative free cash flow of $5.9 billion in the second quarter driven by our investments in CapEx. Free cash flow was $53.3 billion for the trailing 12 months. We ended the quarter with $242.5 billion in cash and marketable securities which includes $87.1 billion of marketable equity securities. Long-term debt was $98.2 billion. And as mentioned in our press release, our Board of Directors declared a quarterly cash dividend on our common stock of $0.22 per share, which is payable in September. Turning to segment results. Google Services revenues increased 15% to $94.5 billion reflecting strong growth in Search and subscriptions. Total advertising revenues were up 14% with growth across all major verticals. We experienced strong ad growth related to the World Cup, particularly in YouTube ads. Google Search and Other advertising revenues increased by 17% to $63.3 billion with retail and finance driving the largest contributions. YouTube advertising revenues increased 13% to $11.1 billion, driven by direct response advertising as well as brand with strength in the living room, as Philipp mentioned earlier. Network advertising revenues of $7.3 billion were down 1%. Subscription platforms and devices revenues increased 15% this quarter to $12.9 billion due to strong growth in both YouTube subscriptions, particularly YouTube Music and Premium, and Google One, which was driven by demand for AI plans. Google Services operating income increased 20% to $39.5 billion and operating margin was 41.8%. The Google Cloud segment delivered outstanding results in the second quarter, driven by our Enterprise AI products and services. Cloud revenues were up 82% to $24.8 billion driven primarily by GCP, which grew faster than Cloud overall; core GCP, AI solutions and AI infrastructure were all important drivers of growth. We also began to recognize revenues from TPU system sales, which we delivered to customer data centers for the first time in Q2. Cloud revenue growth accelerated meaningfully even after excluding the impact of TPU system sales. Cloud operating income was $8.8 billion, more than tripling year-over-year and operating margin increased from 20.7% in the second quarter last year to 35.6%. Google Cloud's backlog increased by more than $50 billion sequentially, reaching $514 billion in the second quarter. The increase was driven by strong demand for our Enterprise AI offerings. The majority of the backlog is related to typical GCP contracts for a broad mix of customers, and we expect to recognize just over 50% of the total backlog as revenue over the next 24 months. In Other Bets, revenues were $382 million, and operating loss was $1.8 billion as we continue to invest in Waymo's business and other key bets. And at the Alphabet level, the operating loss was $5.8 billion, driven by shared AI R&D expenses. Turning to our outlook and business performance expectations in the second half of 2026. First, in terms of revenues, at the current spot rates, we would expect a slight FX headwind to our consolidated revenue in Q3 compared to a 1 percentage point FX tailwind in Q2. This impact will be seen primarily in Search and YouTube ads. In Google Services, as Philipp mentioned, we continue to benefit from innovation in Search by improving the user and advertiser experience with AI. At the same time, in Q3, we will begin lapping an acceleration in Search performance that began in the third quarter last year. In Google Cloud, we're seeing significant demand for products and services which we expect to drive strong growth. As I mentioned earlier, we started delivering TPU systems to customer data centers in the second quarter. We continue to expect to recognize a relatively small portion of the revenues from our existing TPU system sales agreements this year, ramping as we exit 2026. We anticipate the vast majority of the revenues from these agreements will be realized in 2027. And given the supply constrained environment, we plan to expand the use of third-party capacity in Q3 as a bridging strategy while we build up more internal capacity. This strategy allows us to keep growing our customer base and capture greater overall value. However, it will create modest margin pressure in the near term as we utilize this capacity. Moving to investments. We are updating our full year 2026 CapEx guidance range to $195 billion to $205 billion, up from our previous estimate of $180 billion to $190 billion. The increase in the range is primarily due to an acceleration in the delivery of capacity to meet growing demand. As we previously shared, we continue to expect our CapEx to increase significantly in 2027 and we'll provide more details at a later date. In terms of expenses, the significant increase in our investments in technical infrastructure will continue to put pressure on the P&L in the form of higher depreciation expense and related data center operations costs such as energy. We also expect to continue hiring in key investment areas such as AI and Cloud and we are investing in marketing to support our AI products. And finally, we expect free cash flow will remain under pressure, driven by our investments in technical infrastructure which enables us to capitalize on the AI opportunity and continue to drive attractive returns. Q2 represented another strong quarter. Our teams continue to deliver impressive innovation, executing with discipline and velocity. I want to take this opportunity to thank our employees for their contribution to our performance. Sundar, Philipp and I will now take your questions.
Questions and answers
Our first question comes from Brian Nowak with Morgan Stanley.
I have two, one for Sundar and one for Anat. Sundar, with more time in GenAI products and tools in the market and investment continuing to step up, can you just kind of give us some perspective on how your view on the size of the overall GenAI ROIC opportunity and the timing of the ROIC has changed now versus one year ago? And then Anat, you talked quite a bit about supply capacity constraints. Maybe a similar question for you around forward CapEx. How is your budgeting philosophy and these constraints you're putting on the forward spend? How are those changing? Do you sort of think about the right amount to spend in 2027 to close this capacity constraint?
Look, thanks, Brian. I mean it's a good question. I do think it feels like we are in the very early innings of what feels like a secular shift across multiple areas in our core information businesses. Just the possibilities — when I see what all you can do with the absolute frontier capabilities — there's still a lot of work ahead to translate all that into experiences for our consumer users. You can think about end-to-end agentic experiences to really meaningfully do a lot more for them. So all of that looks like extraordinary opportunities with extraordinary returns for executing well on those opportunities. Similarly, on the enterprise side, as you can see from our demand, which is reflected in our growth rates, et cetera — again, in my conversations with many CEOs, many companies, they're all still barely scratching the early stages of what's possible here. So I think we used to talk about Cloud itself — very small percentage of overall workloads in enterprises had shifted to cloud. Now think about what percentage of workloads are really AI-native and AI-enabled. It, again, feels like very, very early. So I would say from an ROIC standpoint, I think we are taking a full-stack approach. We are seeing momentum across consumers and enterprises and developers and so on. So it feels like, if anything, over the past year, we've gotten more bullish on the opportunities ahead. I'll turn it over to Anat.
How do we think about forward-looking CapEx in the context of supply constraints? So we're still in a supply-constrained environment. I think we've said this now for multiple quarters in a row. And we are seeing very strong demand, both from external cloud customers as well as across the business. Our goal is to invest as long as we see an attractive return on that investment, as Sundar mentioned earlier. We do take a long-term view. So we take a multiyear view at what the needs are as well as focus on next year and the near term and building aggressively to meet those demands. And as you've seen, while we have increased our capacity quite significantly over the past three years, the demand still outpaces that investment. And we are, like the rest of the industry, working in a supply-constrained environment. So we're working hard to do that. We do have the benefit of having the full-stack approach. So we're able to drive operational efficiencies and technological efficiencies within our technical infrastructure organization so that we can deliver more compute. But as long as we see these attractive opportunities to invest, we will continue to invest.
Our next question comes from Doug Anmuth with JPMorgan.
One for Sundar and one for Anat. Sundar, can you just talk about your confidence level that the Gemini models can remain at the frontier? I know you had the Gemini Flash model releases this week. But Google perhaps doesn't make maybe quite as much noise as some of the other leading labs or perhaps have quite the frequency. Just curious how you'd address any concerns about that ability to continue developing leading models? And then just related, can you talk about your plans for coding and how you look to close the gap in that part of Enterprise. And then, Anat, not just following the recent equity and debt raises, can you just help us understand how you think about optimal capital structure and how you view the cost of debt versus equity?
Thanks, Doug. Look, I think the frontier is an incredibly dynamic space, and it's fiercely moving forward. At any given moment, when you take a snapshot, it feels dynamic. We have clearly had frontier models. There are many attributes on which we are still at the frontier and there are areas where we have acknowledged we need to improve — coding is an example of that — and I think the teams are very, very focused on it. With 3.6 Flash — and by the way, Flash is our workhorse model — we see tremendous demand for it given it generally hits a sweet spot of performance, cost, reliability and latency. We've incorporated Gemini Flash particularly across our entire portfolio of solutions, be it in cybersecurity, data analytics. If you think about an area like customer service, you need really good voice quality, you need live streaming, you need the ability to reason on that. Or if you're a professional services firm, you need high-quality summarization, content generation, etc. So Flash does very well on all of that. In terms of agentic coding, we are iterating, and you will see us make continued iterations. 3.6 Flash, for example, compared to 3.5 Flash jumped more than 10 points on DeepBench-style benchmarks — and it is more token efficient doing so. So we are clearly using it internally. We are testing it with many customers in coding. So we see the progress just in six weeks from the prior version to the new version, and you will see more continued iterations on that as well. In terms of the frontier, we are both very committed and very confident about being at the frontier. For the next generation of frontier models, you're going to need much larger base models. We are now training Gemini 4, and we're being very ambitious with it. I am very excited by the progress I'm seeing internally on Gemini 4, and I'm confident that people will be pleased when we put it outside. But we will need Gemini 4 as a larger base model to compete at that frontier level. And so we are focused on executing on that well.
And on the question regarding our capital structure and recent equity and debt raises. As we think about our investment needs — and as I said, we look at the next year and multiple years out — we first look at how much we can support based on cash from operations. And as you've seen in our results today, we continue to generate very healthy strong cash flow from operations. So that's our first source of funding. Then we look at debt and most recently we did the equity raise. We have expanded our debt portfolio quite significantly over the past 12 months. If you look a year ago, we were at about $16 billion, and we went to about $100 billion across multiple currencies and geographies. So we're expanding that portfolio. But we also want to make sure we have a resilient and strong balance sheet, which is the rationale behind expanding into the equity markets. At this point, we're not planning to go back to the equity markets with the exception of, as you recall, part of our equity offering was the ATM or at-the-market offering that we will do to address the tax on stock-based compensation, which we'll do for some period of time. But we look at those three dimensions — cash flow from operations, debt levels and equity — in a way to balance it to ensure that we continue to maintain a healthy balance sheet.
Our next question comes from Eric Sheridan with Goldman Sachs.
Maybe two, if I could, both on TPUs. Sundar, can you talk about some of your key learnings as you scale your TPU efforts in terms of both the demand for TPUs and how you think about balancing the external demands for TPUs versus the internal demand for custom silicon inside the organization as you look over the next couple of years? And then Anat, if I could just follow up. You made some comments in your prepared remarks about the impact of TPUs on Google Cloud. I didn't know if you could give us a little bit more granularity about the size of TPU system sales as a part of the revenue backlog and how to think about the revenue recognition and the potential for margin impact from TPUs in the years ahead.
Okay. In terms of our TPU roadmap progress and the value in terms of performance and the edge it gives, we are very pleased. We obviously use it extensively. In terms of allocating TPUs, our first priority is making sure we are allocating what we need to compete at the frontier in terms of AGI development, and that is the foundation for everything we do. But given the extraordinary demand, we balance external demand even for most cloud customers. For serving our models, think Vertex AI and Gemini Enterprise — the momentum we see — we are using both TPUs and GPUs, mainly for serving our models. So we're using them for those purposes. To the extent people want TPUs as infrastructure, we are balancing it by increasingly looking at opportunities to put TPUs in their data centers or in other data centers via partnerships, like the project we are doing with third parties. So we're using that as an ability to balance and make sure we are allocating as much of the available compute for frontier model development as well as serving both our consumer business and our enterprise business in terms of first-party models.
And in terms of revenue recognition and how we look at the TPU system sales, the way to think about it is the following. When we signed the agreements that I mentioned in the prepared remarks, they are then reflected in the cloud backlog. Now the vast majority of the $514 billion of cloud backlog is GCP agreements, but the TPU system sales are reflected in that backlog. We start building inventory to be able to sell those systems. So you see that impact on cash from operations because we built ahead, obviously, as we're building that business and ramping up. And then once we start delivering the systems, that's when we start recognizing revenue. This quarter was a small amount of the total, and that total agreement will continue to ramp up throughout 2026, but then you'll see the vast majority of the revenue from that agreement come through in 2027.
Your next question comes from Ross Sandler with Barclays.
Just back to the AI model war that's going on, Sundar. Just wanted to ask about the speed of model releases and how you can potentially speed up the cadence by which Gemini is releasing models. So we all saw the third-party compute deal with SpaceX. What else are we doing to kind of speed up the pace of model releases compared to some of the other folks out there? And then you just mentioned the Flash models, which have been very successful, to your point. Do you view that as the right end of the market to be in, given how crowded the lower-cost end of the model space has become? Just any thoughts on how you see all that playing out?
Thanks, Ross. Two things I'd say. First of all, we really think about the Pareto frontier. We want to make sure for our customers we are offering the best models at various price points. So you will see us across that spectrum. But it's very important to us to have the best frontier models out there as well as models that are very performant and low cost. This is why we have Flash, Flashlight, Flash Pro, etc. On the speed of model releases, I do think you will see us continue to pick up pace. Obviously, you've seen us release models: we announced 3.5 Flash at I/O, we have followed that up with 3.6 Flash and you'll see continued iterations of that model with further progress in agentic coding, etc. We are really focused, putting a lot of effort into Gemini 4. It's a very ambitious effort. We want to compete at the frontier level of where the frontier will be when Gemini 4 comes out. So we are applying a lot of our compute and effort in that direction. But on top of that, we are creating a baseline on which you will see us rapidly iterate subsequent model releases. Picking up pace and releasing models on a faster cadence is part of our roadmap as we build Gemini for scale as well.
Your next question comes from Michael Nathanson with MediaBin.
Sundar, just to keep up with the model theme of this call, can you talk a bit about what do you think the moats are to this new model era or what are your strategic advantages that fuel your business that even if everyone gets to the same type of model capabilities, what gives you the edge to keep growing like you're doing? And then not back to Eric's question, if you could talk about the margin impact on TPUs going forward, is it incremental margin to the core cloud margin or is it a lower margin from the group?
Look, Michael, there are many layers at which we are providing solutions. I think part of what the value of our full-stack approach is, customers are coming to us for solutions. So if you take an area like cybersecurity or data analytics, people are deploying solutions in which models are an ingredient in that solution. Take cyber: people are using Chronicle and Virus and with our upcoming tools, they're using it to detect and patch vulnerabilities. Or take data analytics: people are bringing what has been in the past siloed data sources together and putting an intelligence layer on top with Gemini Enterprise. So the model is just an ingredient in those solutions. It's important to remember that. Even where you're purely on a model consumption basis, models are increasingly becoming end-to-end orchestrated systems — workflows, agentic workflows. To develop all this, it's your ability not just to bring compute, but you need to train and serve, data quality, the environments you have, your ability to continually improve, provide customers with peace of mind that their data is confidential to them and does not flow back to the shared models. Their ability to configure and serve all this securely, manage it and provision it — these are big end-to-end things. This is what our Cloud business as a whole is focused on building, and we are seeing demand across all of these components. Having our own models allows us to optimize these solutions and bring an integrated offering, though we will also provide other models as part of these solutions. We also bring infrastructure as part of the solutions and take a full-stack approach. So I think we are very well positioned there.
On the TPU margins, we don't break out margins for any specific products or infrastructure components. There are benefits from designing and manufacturing our own chips. But the way to think about it is this expands our total addressable market. It expands the opportunities by providing solutions to customers that need those systems in their data centers. Overall, if you think about Cloud margins, we had a fantastic expansion of margin to 35.6% this quarter — outstanding operational discipline and leveraging top-line growth. Now as you think about Q3 and the remainder of the year, given the supply constraint environment, we are planning to expand the use of third-party capacity in Q3 as a bridging strategy while we continue to build out internal capacity. That capacity, given the cost of that third-party capacity, will put some pressure on operating margins for Cloud in the near term. And the integration of Wiz does create some near-term headwind related to the acquisition.
Your next question comes from Mark with Bernstein.
And just on these third-party deals and kind of the bridge in capacity, is there a particular objective where the constraints are most severe? Or is it really broad-based? And I guess stepping back a bit and thinking about capacity allocation across the businesses, has that changed at all? Or is there a way to contextualize how you're now thinking about the trade-off between allocating compute to Search versus model training versus for the Cloud business?
Mark, I can comment on it. The baseline with which we start is what it takes to continue AGI development at the frontier. The priors on that are just based on where the model frontier moves. So we start with that as a baseline. Beyond that, we prioritize our core product areas like Search and YouTube as well as Cloud. Within Cloud, we prioritize the compute to make sure we can serve our models in the context of Vertex and Gemini Enterprise and our core solutions, be it data analytics and cybersecurity. So our core serving for our core products across consumers and enterprise is where the compute is primarily going, and that's how we think about it. On the bridge deals, the main thing I would say is these are very large customers on Cloud who we are trying to support through this extraordinary moment. The incremental opportunities they are bringing to us — while the short-term cost over a few months may be high — in the lifetime of the deal as we bring more capacity on is highly ROI positive. So those are factors we take into account. Are you willing to take an upfront short-term cost to support a customer in what is a multiyear opportunity where the margins and returns are very attractive? Those are the considerations.
Your next question comes from Ron Josey with Citi.
Maybe I'll switch topics somewhat, and Philipp, ask you a little bit more about monetization from a Search and YouTube perspective, just given the greater signal we have with AI searches. You talked about YouTube strength. Can you talk to us just a little bit more on how advertisers are leveraging the greater personalization and targeting that Google has to drive this growth? And a question we often get is, Google is that size and scale and we're still growing 17%. Any reasons what's driving that all-time high on searches?
Thank you so much for the question. Our 17% year-over-year growth rate in Q2 Search and Other revenue was really driven by many parts of our business working well together and very deep Gemini integration. All the major verticals contributed to the growth: retail was the greatest contribution followed by meaningful contributions from finance, tech, media, entertainment and so on. Gemini supercharges our ability to understand what people are looking for and match the right ad. We're applying Gemini models across our entire ad infrastructure — ad quality, advertiser tools, and new AI experiences — and we're deeply integrating Gemini into customer tools to make campaigns more efficient. On top of this we have AI-first campaigns like AI Max that help advertisers adapt and find opportunities beyond keywords. That ability for us to go deeper and target better and AI Max continues to unlock searches that weren't monetizable before. On the YouTube side, ads growth was driven by direct response and brand advertising. We continue to see a lot of momentum in the living room and an exciting ads roadmap ahead on both brand and direct response sides.
Your next question comes from Ken Goralski with Wells Fargo.
Two questions, please. First, how would you assess the forecasted returns on compute capacity investments in '27 compared to prior years '25 and '26 in light of the supply chain constraints and supply chain inflation we're seeing? I'm curious as to how you think about the return profile of '27 in kind of future investments versus past investments in capacity? And the second one, a little different topic here. When you think about Waymo, what are the key factors you would consider when evaluating a change in corporate structure for Waymo? I know you've previously been reticent to talk about this, but the business is scaling clearly, there's leadership there and has a lot of momentum. Under what conditions would it make sense for Waymo to live outside of Alphabet?
Ken, on the first question, we are working with a disciplined ROIC framework. To the extent input costs change, we reflect that in our ability to price our solutions and see returns. Our compute capacity investments for 2027 are planned based on strong demand indicators, including long-term deals and renewals with exceptional demand. We are using all that to plan and invest accordingly. If anything, the dynamics look healthier than they did about a year ago, which gives us confidence to undertake those investments. On Waymo, we are really focused on executing and scaling Waymo. We've set it up with the bet structure and have given the team a lot of support. Alphabet continues to plan and invest long term in the business. Our ability to think and plan long term and give them the confidence to undertake a long-term roadmap and scale the business is hugely valuable. Right now we are focused on scaling the business and executing on that potential.
Your last question comes from Sridat Kajuria with Wolf Research.
Could I please ask two? One is on TPU long-term strategy. Longer term, is it fair to think that as TPU sales scale you'd build a merchant silicon business with a software stack that goes with it? And second is on YouTube revenue: given how strong engagement is, what are some of the factors that could drive an accelerating growth rate at YouTube on a go-forward basis?
Maybe I'll answer the first one, and Philipp can broadly talk about YouTube revenue growth as well. On TPUs, we've been operating with TPUs for a while. A lot of our Cloud serving is done on TPUs, and we are also seeing strong demand for TPU stand-alone systems from other customers. We'll scale based on the opportunities we see and the demand commensurate with the constraints that exist and the allocation needs we have for frontier model development, making sure we can serve our consumer businesses and enterprise businesses well. I don't want to project too far into the future, but we will scale up based on demand and opportunity.
Yes. On YouTube, I mentioned several drivers of ads revenue growth. We're excited about our ads roadmap. For brand, we see ongoing opportunities on Connected TV and the living room. We're building tools to help advertisers find creators at scale across screens and then measure the results. We have a slate of new and returning creator shows that are coming exclusively to YouTube. We see opportunities in direct response, particularly with Demand Gen and Shorts; these remain interesting opportunities to expand our base to more small and medium advertisers across more verticals. Shorts creates more opportunities for less disruptive ads, increasing overall ad effectiveness. We will continue to innovate in direct response with shoppable ad formats in the living room, which drives retail strength. We announced Buy with Google Pay enabling CTV viewers to complete purchases directly on their TV with two clicks, which is a compelling long-term opportunity. We launched the affiliate partnership Boost to drive incremental sales and greater earnings from YouTube shopping affiliate commissions. Overall, very exciting times for monetization on YouTube.
And that concludes our question-and-answer session for today. I'd like to turn the conference back over to Jim Friedland for any further remarks.
Thanks, everyone, for joining us today. We look forward to speaking with you again on our third quarter 2026 call. Thank you, and have a good evening.
Thank you, everyone. This concludes today's conference call. Thank you for participating. You may now disconnect.