Google Bets $40 Billion on Anthropic as AI Compute Race Escalates

Alphabet will invest up to $40 billion in Anthropic, the company behind the Claude AI model family. $10 billion lands in cash now at a $350 billion valuation, with another $30 billion contingent on Anthropic hitting performance targets, according to Reuters. Bloomberg first broke the story.

The deal comes just days after Axios reported that Amazon announced its own investment of up to $25 billion in Anthropic. Together, the two tech giants are committing up to $65 billion to a company that competes directly with both of them in the AI model market.

Anthropic confirmed the deal in a statement: Google is committing $10 billion in cash at a $350 billion valuation to help support a major expansion of computing capacity. The additional $30 billion will follow if Anthropic meets performance milestones.

The strategic logic is straightforward even if it looks paradoxical. Google is both competitor and critical infrastructure supplier to Anthropic. The $40 billion deal is not charity — it is a compute play.

As TechCrunch reported, Anthropic relies heavily on Google Cloud for TPUs — Google's specialized AI chips that are among the best alternatives to Nvidia's processors. Every dollar Anthropic spends on compute through Google Cloud generates revenue for Alphabet. The investment creates a flywheel: Google funds Anthropic, Anthropic buys Google Cloud capacity, Claude gets better, demand grows, Anthropic needs more compute.

The deal also gives Google equity upside. Anthropic's valuation has skyrocketed from $4.1 billion in early 2023 to $380 billion as of its February 2026 Series G round, with VC offers reportedly valuing it at up to $800 billion. If Anthropic goes public — reports suggest an IPO as early as October 2026 — Google's stake could be worth multiples of its investment.

The investment is structured as cash plus compute access. Here is how the pieces fit together:

The AI competition has shifted from who has the smartest model to who has the most compute. Interesting Engineering frames it bluntly: capital alone is not enough — access to reliable, large‑scale computing power has become the real battleground.

OpenAI made this explicit in an investor letter in early April, boasting about its outsized compute access. The company projects reaching 30 gigawatts by 2030, compared to Anthropic's estimated 7‑8 gigawatts next year. OpenAI co‑founder Greg Brockman said on The Knowledge Project podcast: "We are moving to a compute‑powered economy."

But Anthropic is closing the gap fast. With the Google deal adding 5 GW, the Broadcom partnership contributing 3.5 GW starting in 2027, and Amazon providing nearly 1 GW by year‑end, Anthropic's total compute pipeline is scaling aggressively. The question is whether it is fast enough.

The investment is being justified by extraordinary growth. Anthropic's annualized revenue has surpassed $30 billion as of April 2026, according to Anthropic's own announcement — up from approximately $9 billion at the end of 2025. That is a 3x increase in roughly four months.

The growth is driven heavily by Claude's coding tools. Claude Code, Anthropic's agentic coding assistant, has become a major revenue driver, with estimates putting its run‑rate at $2.5 billion+ by early 2026. The tool competes directly with GitHub Copilot and Cursor in the AI coding market, where the three now control over 70% of the market combined, according to Brookings Institution analysis.

Enterprise adoption is also accelerating. When Anthropic announced its Series G in February, over 500 business customers were each spending more than $1 million annually. By April, that number had doubled to 1,000+ customers, according to Anthropic's blog post.

For developers and companies building on the Claude API, the Google investment carries several practical implications:

Capacity relief is coming, but not immediately. The new Google Cloud capacity (5 GW over 5 years) and Broadcom TPU deal (3.5 GW starting 2027) will ease the capacity constraints that have periodically caused rate limits and availability issues. But most of this capacity comes online in 2027 and beyond — builders should not expect instant relief.

Multi‑cloud availability is now a genuine procurement advantage. Claude remains the only frontier AI model available on all three major clouds: AWS Bedrock, Google Vertex AI, and Microsoft Azure Foundry. With Google deepening its partnership, Vertex AI's Claude integration will likely get priority support and potentially better pricing for enterprise buyers who already use Google Cloud.

The competitor‑partner dynamic creates risk. As Brookings documented, Anthropic has a history of cutting off customers that compete with its own application layer — it blocked Windsurf when OpenAI tried to acquire it, cut off OpenAI's Claude access, and blocked xAI. Any builder creating tools that overlap with Anthropic's own products (coding, legal, financial) faces the risk of being deplatformed.

API pricing may stabilize. With more compute coming online and revenue growing 3x, Anthropic has less pressure to raise prices aggressively. The competitive dynamics with OpenAI and Google's own Gemini model also keep pricing in check.

The investment news comes at an awkward time for Anthropic. Days before the announcement, BBC reported that unauthorized users gained access to Claude Mythos, Anthropic's most powerful model designed for cybersecurity applications. The model is restricted because of its ability to discover and exploit vulnerabilities at scale.

According to the report, a small group accessed Mythos through a third‑party contractor's shared credentials — not through a direct hack, but through what cybersecurity experts described as misuse of existing access. Anthropic stated it is investigating and has found "no evidence that the supposedly unauthorized activity has impacted Anthropic's systems," as TechCrunch reported

Richard Horne, head of the UK's National Cyber Security Centre, warned at the CyberUK conference that "frontier AI is rapidly enabling discovery and exploitation of existing vulnerabilities at scale," as DIGIT reported The breach raises questions about Anthropic's ability to control access to its most dangerous systems even as it scales rapidly — a concern that builders relying on Claude for sensitive applications should monitor closely.

The Google‑Anthropic deal is a milestone, but it is more waypoint than destination. Three big questions remain:

Will regulators intervene? The UK's Competition and Markets Authority already investigated Google's earlier $2 billion Anthropic investment. A $40 billion commitment will attract far more scrutiny. U.S. Senators Warren and Wyden have already launched investigations into Big Tech partnerships with AI developers. The DOJ is simultaneously pursuing an antitrust case against Google. Adding a $40 billion investment in a key competitor‑partner complicates that picture significantly.

When does Anthropic go public? Reports point to an IPO as early as October 2026, though FutureSearch forecasts a more conservative March 2027 timeline at a median valuation of $560 billion. The Google investment effectively serves as a pre‑IPO round, locking in a major strategic shareholder.

Can Anthropic close the compute gap with OpenAI? Even with Google's 5 GW and Amazon's capacity, Anthropic trails OpenAI's ambitious 30 GW target for 2030. The compute race is far from over. As OpenAI's investor letter argued, having the best model matters less than having the capacity to serve it at scale — and that is the game Anthropic is now playing with house money. Lots of house money.