Alphabet (GOOGL) Q1 FY26 earnings call: guidance and Q&A

Given compute constraints, what areas in Search will receive next-gen compute for ROIC, and what is the philosophical strategy for pricing TPUs sold to third parties?

Sundar highlighted agentic workflows in Search as a huge opportunity for next-gen compute. For TPUs, the strategy is to help Google Cloud customers, including direct sales to select clients like capital markets firms and AI labs, while ensuring overall ROIC and economies of scale for Google's compute environment.

“But our investments in our full stack of AI approach, I think, puts us in a good position to bring those experiences to search and I'm pretty excited about it.”

asked by Brian Nowak · answered by Sundar Pichai

How does Alphabet think about the increasing CapEx trajectory, especially for 2027, and its ability to service the massive Cloud backlog?

Anat stated that CapEx increases are thoughtful, driven by robust demand and the opportunity ahead. The company aims to capitalize on this demand responsibly, noting the strong ROIC seen in Cloud and Search growth. More clarity on 2027 CapEx will be provided later.

“So as we're seeing that robust demand across the business, we are looking at what can we do to support that growing demand and the opportunity ahead of us and increasing CapEx to meet that demand.”

asked by Douglas Anmuth · answered by Anat Ashkenazi

How does Alphabet's unique AI infrastructure position it competitively, and what does the Universal Commerce Protocol (UCP) mean for the services business as agentic commerce scales?

Sundar emphasized Alphabet's differentiation through its vertically optimized AI stack, owning frontier models and silicon, and deep security investments. Philipp explained UCP as an open standard for agentic commerce, co-developed with industry leaders, designed to remove shopping friction and enable new checkout experiences in AI Mode and Gemini app.

“I do think we are genuinely differentiated. We are unique in the market because of our vertically optimized AI stack and the way we codevelop the components from our infrastructure and models to platforms and the tools to applications and agents.”

asked by Eric Sheridan · answered by Sundar Pichai

How does Alphabet allocate constrained compute capacity across internal and external projects, and what are the plans for advertising on the Gemini app?

Sundar explained that compute allocation prioritizes R&D for frontier model development (GDM), followed by long-range plans for core areas like Search, YouTube, and Google Cloud, all within a robust ROIC framework. Philipp stated that Gemini app monetization currently focuses on the free tier and subscriptions, with ads potentially coming later, as ads can be valuable commercial information.

“Obviously, we are compute constraint in the near term. And as an example, our cloud revenue would have been higher if we were able to meet the demand.”

asked by Michael Nathanson · answered by Sundar Pichai

How do verticalized capabilities help navigate supply chain complexities and inflation for CapEx, and are there future use cases where Search monetization shifts to subscriptions?

Sundar noted that Alphabet's scale and full-stack approach foster deeper partnerships across the supply chain, factoring in inflation for CapEx commentary. He added that while Search is largely ad-supported, valuable use cases may lead to subscription monetization for powerful models, citing momentum in AI subscriptions for better Gemini models.

“But the future, as you are right, that in a valuable as we serve more and more valuable use cases, there are going to be use cases where people will want to use the most powerful model. And there may be different ways to accomplish that.”

asked by Kenneth Gawrelski · answered by Sundar Pichai

How should investors think about the TPU sales opportunity, the breakdown of backlog growth between TPUs and cloud, and the margins associated with large generative AI cloud deals?

Sundar highlighted tremendous demand for both AI solutions and infrastructure, including TPUs and GPUs, allowing Google to meet diverse customer needs. Anat clarified that while TPU hardware agreements are in the $462 billion backlog, the majority is still GCP agreements. Sundar reiterated that all allocation decisions, including large AI deals, are made within a robust ROIC framework.

“Look, overall, I would say, look, we see tremendous interest in -- there's tremendous demand for both AI solutions as well as AI infrastructure, including massive interest in our GPU offerings as well as TPUs.”

asked by Justin Post · answered by Sundar Pichai