Google Bets Big on AI Chip Race

Google is making a substantial play in the escalating AI compute race, planning a significant $40 billion investment in Anthropic. This move underscores Google's commitment to expanding its AI capabilities and infrastructure.

Mizuho estimates that approximately $61 billion of Google's cloud backlog, extending through 2027, could originate from sales of its Tensor Processing Units (TPUs). A substantial portion of this revenue is anticipated to be recognized in the upcoming year, offering investors an alternative avenue to participate in the AI hardware market.

This strategic investment comes as Google projects capital expenditures to reach up to $190 billion this year, more than doubling its spending from 2025. To ensure a return on this immense investment, Google must demonstrate sustainable revenue generation from the broader AI ecosystem and clearly articulate its agent strategy with Gemini at the upcoming Google I/O event.

Analysts acknowledge the risks associated with Alphabet's significant capital expenditure but also view its capacity to fund these outlays as a competitive advantage compared to rivals like OpenAI. The company's transition from an AI laggard to an infrastructure leader has been rapid, but continued investor confidence hinges on successful execution and innovation.