Alphabet Q1 Earnings Call: AI Agents to Transform Shopping, TPU Sales to Clients

Alphabet Inc., the parent company of Google, reported its financial results for the first quarter ended March 31, 2026.

Following the earnings release, Alphabet CEO Sundar Pichai, Chief Business Officer Philipp Schindler, and CFO Anat Ashkenazi hosted an analyst conference call to discuss the results.

**Morgan Stanley Analyst Brian Nowak:** I have two questions. First, for Sundar: You recently mentioned significant constraints on compute resources. Where do you see the highest potential return on invested capital for Search from next-gen compute over the next 12 months? Second, regarding selling TPUs to third parties: Can you explain the pricing strategy, considering the high ROIC from using TPUs for multi-year Google Cloud workloads?

**Sundar Pichai:** On your first question about Search, we are leveraging our investments in Gemini models to drive innovations like AI Overviews and AI Mode, which are contributing to increased product usage. Looking ahead, a significant opportunity lies in integrating intelligent agent workflows into both Search and the Gemini app in a user-friendly way. We are well-positioned due to our full-stack AI strategy.

Regarding TPUs, our framework is about helping customers through Google Cloud. In certain cases, it makes sense to provide TPUs directly to clients, such as those in capital markets or advanced AI labs running high-performance workloads in their own data centers. We will pursue these opportunities while focusing on the broader Google Cloud potential. A large part of the cloud opportunity involves providing infrastructure, and sometimes directly selling DPU hardware, all guided by an ROIC-focused approach that also supports economies of scale for our R&D.

**JP Morgan Analyst Doug Anmuth:** I have questions for Anat and Philipp. Anat, with the significant increase in capital expenditure expected for 2027, how does the current growth trajectory support the substantial backlog? Philipp, what is driving the record-high search query growth, and what is the potential to increase the ~20% query coverage by ads?

**Anat Ashkenazi:** On capital expenditure, we have increased CapEx annually, making prudent decisions to meet internal and external demand. The ROIC is validated by growth in Search, Cloud, and the opportunities in the cloud backlog. We will support this growing demand responsibly and provide more specific figures in future calls.

**Philipp Schindler:** We are very pleased with the advertising performance. Strong search results are driven by multiple factors, not a single one, with contributions from retail, finance, and all major industries. We make hundreds of quarterly adjustments to improve user and advertiser experiences. The introduction of new AI features has not hindered ad performance. Search query volume is at an all-time high, fueled by AI Overviews and AI Mode, which also boost commercial queries. The 20% ad coverage ratio has room to grow, aided by AI's better understanding of user intent and Gemini's ability to monetize longer, more complex queries. Deploying Gemini across our ad infrastructure is improving ad quality, advertiser tools, and the AI user experience.

**Goldman Sachs Analyst Eric Sheridan:** First, how does your AI infrastructure strategy help you scale compute capacity while maintaining margin and efficiency advantages? Where do you stand competitively? Second, for Philipp: What does the Universal Commerce Protocol (UCP) mean for the services business as agentic commerce develops?

**Sundar Pichai:** We have a differentiated advantage through our vertically optimized AI stack, which integrates infrastructure, models, platforms, tools, apps, and agents. Our frontier models and custom silicon help us maintain leadership, complemented by deep investments in security. We think about this through the lens of Google Cloud and can meet customer needs in ways others cannot. Continued investment keeps us at the forefront.

**Philipp Schindler:** Our decisions are based on real demand signals and an ROIC framework. On UCP, agentic commerce is transformative, changing shopping from discovery to decision-making and helping brands differentiate. It aims to remove friction, so consumers no longer have to choose between speed and certainty. UCP is a new open-source standard developed with partners like Shopify, eBay, and Walmart. It has received positive feedback from hundreds of top tech companies, payment partners, and retailers. UCP will support new checkout experiences in Search and the Gemini app, allowing shoppers to research and buy directly from merchants on Google. We are optimistic about its potential.

**Barclays Analyst Ross Sandler:** As agentic shopping becomes a reality, how will the shift from traditional ads to this new, frictionless model impact core advertiser pricing and volume trends, especially in Search?

**Philipp Schindler:** Our primary focus is the user experience. By designing agentic workflows that show users valuable information beyond just price, we create a foundation for exploring richer ad formats. This includes AI-enhanced experiences like our virtual try-on tool and Google Lens. Focusing on the user experience will naturally lead to improvements in other areas.

**MoffettNathanson Analyst Michael Nathanson:** For Sundar: With compute constraints, how do you prioritize allocating excess capacity between internal and external projects? What criteria are used? For Philipp: With more images appearing in shopping journeys on the Gemini app, are you considering introducing ads there, and what factors would influence that decision?

**Sundar Pichai:** The foundational principle is the compute required for R&D and training frontier models. We create long-term plans for core areas like Search, YouTube, and Google Cloud. Demand is very strong, evidenced by multi-fold year-over-year growth for enterprise AI solutions this quarter, including demand for Gemini Enterprise, cloud infrastructure, and, in some cases, TPU hardware. We balance all these factors. Current constraints mean our cloud revenue could be higher with more capacity, but we are addressing this through investment and have a solid long-term planning framework for allocating resources to unprecedented opportunities.

**Philipp Schindler:** For the Gemini app, our focus is on creating a great user experience. We believe ad formats that perform well in AI Mode could be successfully extended to the Gemini app in the future.

**Bernstein Analyst Mark Shmulik:** For Philipp: Beyond query growth, are AI tools shortening consumer shopping journeys and improving conversion rates? Can you quantify this behavioral change's contribution to Search growth compared to new advertiser AI tools?

**Philipp Schindler:** The key is that Search is in an expansion phase. AI has fundamentally changed how people search. Query volume is at record highs. We've moved beyond text links to AI Overviews, AI Mode, Google Lens, and Circle to Search, making Search smarter and capable of handling complex queries. AI-driven search campaigns now allow smaller businesses to reach more customers. Combined with tools like Google Translate, we are in a strong position.

**Citi Analyst Ron Josie:** Can you break down the drivers of continued margin expansion, particularly for Cloud, given that AI-related revenue is often thought to have lower margins?

**Anat Ashkenazi:** We are pleased with the margin expansion. Strong revenue growth across segments, including Cloud, provides leverage. We continuously improve operational efficiency across the organization and in our technical infrastructure. Our CapEx investments in data centers and servers, alongside process innovations, contribute to efficiency. Depreciation from these investments affects costs for both Google Cloud and Google Services. Google Cloud's margin improvement is largely due to strong revenue growth and highly efficient operations by the team. We are also using AI to optimize internal operations, like coding assistance with Gemini. We will continue driving efficiency, though higher CapEx will bring increased depreciation, which will pressure margins.

**Wells Fargo Analyst Ken Gawrelski:** First, how does your vertical integration help navigate the complex supply chain, especially with inflation and constraints? Are supply chain price inflation factors considered in your 2026/2027 CapEx outlook? Can you update on the allocation of compute capacity between internal use and external cloud? Second, with new search use cases emerging, could some consumer scenarios be better monetized via subscription, changing the ad-to-subscription mix for the "new search"?

**Sundar Pichai:** On compute allocation, we use long-term planning and an ROIC framework. Our operational scale, diversified business advantages, and relationships with supply chain partners help us navigate the complex environment. Economies of scale are beneficial. For Search, our advantage is developing frontier models while considering capability and cost, allowing us to serve users at scale. As we offer higher-value scenarios, users may want access to more powerful models, potentially through different models. We already offer various subscription tiers for enhanced Gemini access within the Google user experience, including Search. Our AI subscription business showed strong growth this quarter, and this model will help us meet diverse user needs.

**Bank of America Analyst Justin Post:** What is the opportunity seen in TPU sales? Can you break down the backlog between TPU-related and cloud service-related portions? What are the margin profiles of large generative AI cloud deals?

**Sundar Pichai:** Our GPU and DPU products are also important, offering customers a diverse choice to meet their needs.

**Anat Ashkenazi:** To add on the backlog composition: The revenue from TPU hardware agreements mentioned by Sundar is included in the $46.2 billion cloud backlog, but the majority remains related to GCP contracts. Over half of the backlog is expected to be recognized as revenue within 24 months. For TPU hardware sales, a small amount of revenue is expected later this year, with the majority recognized in 2027.

**Sundar Pichai:** We cannot comment on specific contracts, but we evaluate all opportunities through a disciplined ROIC framework, especially in a constrained compute environment.