Google will invest up to $40 billion in Anthropic, the companies confirmed Friday. The deal starts with $10 billion in immediate cash at Anthropic’s latest valuation of $380 billion, with the remaining $30 billion contingent on Anthropic hitting performance milestones, according to Bloomberg, which first reported the agreement.
The Deal Structure
The investment expands on a longstanding infrastructure relationship between the two companies. Anthropic relies heavily on Google Cloud for access to tensor processing units (TPUs), Google’s custom AI chips that serve as an alternative to Nvidia’s processors. Earlier this month, Anthropic announced a partnership with Google and chipmaker Broadcom to secure multiple gigawatts of TPU capacity starting in 2027. A subsequent Broadcom securities filing put that figure at 3.5 gigawatts, according to TechCrunch.
The new agreement adds another 5 gigawatts of Google Cloud capacity over the next five years, with room to scale further.
Google’s total investment in Anthropic prior to Friday exceeded $3 billion, starting with a $300 million stake in 2023 that gave Google roughly 10% ownership, followed by an additional $2 billion later that year. Google reportedly held a 14% stake before this latest commitment, according to CNBC.
The Compute Arms Race
The deal arrives as Anthropic scrambles to address what it has called “inevitable strain” on its infrastructure from surging demand for Claude. The company faced widespread complaints about Claude usage limits in recent weeks and published a technical post-mortem identifying three system changes that degraded performance across Claude Code, Claude Agent SDK, and Claude Cowork.
Anthropic has been stitching together compute capacity from multiple providers. Weeks before the Google deal, Amazon invested $5 billion in Anthropic as part of a broader agreement under which Anthropic is expected to spend up to $100 billion for roughly 5 gigawatts of compute over time, according to TechCrunch. Anthropic also struck a deal with CoreWeave for additional data center capacity earlier this month, according to The Wall Street Journal.
The pattern mirrors OpenAI’s approach. OpenAI has secured multi-hundred-billion-dollar commitments across cloud providers, chip suppliers, and energy companies, including an expanded deal with chipmaker Cerebras worth more than $20 billion this month. OpenAI has also criticized Anthropic for failing to secure enough compute, according to CNBC.
Revenue and Valuation
Anthropic’s annualized revenue has topped $30 billion, according to CNBC. The company’s valuation stood at $380 billion as of its February funding round, though investors have since been eager to back it at $800 billion or more, per Bloomberg. Anthropic is also reportedly considering an IPO as soon as October.
Much of the capital flowing into frontier labs will return as cloud revenue. Google, Amazon, and Microsoft are effectively pre-purchasing demand by investing in the companies that will consume the most compute on their platforms. The dynamic creates a circular funding model: cloud giants inject capital, which gets spent back on their infrastructure, which generates revenue that justifies the investment.
The Competitive Map
Between Google’s $40 billion commitment and Amazon’s $25 billion agreement, Anthropic has locked in over $65 billion in capital commitments from the two largest cloud providers. Google’s position is unusual: it competes directly with Anthropic through Gemini while simultaneously serving as a critical infrastructure supplier. The same tension exists with Amazon, which offers its own Nova models while hosting Claude on AWS.
For enterprises building on Claude, the deals signal that capacity constraints should ease over the next 12 to 18 months as new TPU and GPU infrastructure comes online. For the broader AI market, they confirm that compute access, not model quality, is becoming the primary bottleneck and the primary competitive moat.