Google Bets $40 Billion on Anthropic as Claude Outpaces Gemini

Alphabet will invest up to $40 billion in Anthropic, making it the largest single‑company commitment to an AI startup outside Microsoft’s OpenAI partnership. The deal structure is staged: $10 billion in cash goes in immediately at Anthropic’s existing $350 billion valuation, while the remaining $30 billion is contingent on Anthropic hitting performance targets, according to Reuters.

The commitment brings Google’s total Anthropic stake to approximately $43 billion, cementing a relationship that started with a $300 million investment in 2023. That earlier bet has returned a 70‑fold valuation increase in under three years, The Next Web reports.

The equity grab is only half the story. The deal secures 5 gigawatts of Google TPU compute capacity for Anthropic over five years — roughly the peak summer electricity load of metropolitan San Francisco. Combined with the 5 gigawatts from Amazon announced four days earlier, Anthropic now has 10 gigawatts of dedicated compute across two independent supply chains, according to Tech Insider.

That’s significant because Anthropic trains on three chip platforms simultaneously — Google TPUs, Amazon Trainium, and Nvidia GPUs — reducing single‑vendor risk in a way that CNBC notes OpenAI’s Azure‑dependent architecture does not. For builders, this means Claude API capacity constraints (remember the late‑2024 Claude 3.5 Sonnet shortages) should ease through H2 2026 as Trillium and Trainium 3 capacity comes online.

Anthropic’s annualized revenue run rate hit $30 billion in April 2026, up from $1 billion in January 2025 — a 2,900% growth rate that ² calls unmatched in American technology history. Claude Code alone generates over $2.5 billion in annual run rate, hitting $1 billion within six months of launch.

Enterprise adoption is accelerating too: Anthropic now counts 8 of the Fortune 10 as customers, with over 1,000 businesses spending more than $1 million per year — a figure that has doubled since February. Reuters reports that Claude’s enterprise LLM API market share sits at 32%, ahead of GPT‑4o’s 25%.

The paradox is deliberate. Google’s own Gemini competes directly with Claude, yet Google is doubling down on Anthropic for three strategic reasons, according to Tech Insider:

• Nvidia hedge: Anthropic is the only frontier lab whose primary training stack runs on something other than Nvidia GPUs. Every additional gigawatt on TPUs validates Google’s silicon as a credible alternative.
• External TPU validator: Internal use cases face “we made it, we use it” skepticism. Anthropic choosing 5 GW of TPUs is the strongest third‑party endorsement possible.
• Pre‑IPO positioning: By committing capital ahead of a potential Anthropic IPO, Google locks in information rights, board observation, and preferred compute pricing impossible to negotiate against a public company.

² puts it bluntly: “You do not invest that sum in a competitor’s product because you are confident in your own. You invest it because the competitor’s product is generating $30 billion in annual revenue and growing faster than anything you have ever built.” Sergey Brin reportedly raised personal alarms about Claude’s superiority in coding tasks.

Four days before Google’s announcement, Amazon expanded its own Anthropic commitment to $25 billion ($5 billion new capital plus $20 billion in conditional follow‑on), per.⁵ The combined $65 billion in hyperscaler pledges in a single week signals that the cloud wars have become what ² calls “AI custody battles.”

The structure is familiar: much of the capital cycles back as cloud spend. Google Cloud distributes Claude through Vertex AI to enterprise customers including Coinbase, Cursor, Palo Alto Networks, Replit, and Shopify. Every API call through Vertex generates cloud revenue for Google regardless of whether Gemini is involved. It’s a cloud computing play disguised as an AI investment.

The deal structure was designed with regulators in mind. Google is contractually capped at 15% ownership with no voting rights, no board seats, and no observer privileges — structured to avoid triggering merger review thresholds, according to The Next Web.

But $43 billion in total investment plus multi‑gigawatt compute dependency creates economic entanglement that no ownership percentage can fully capture. The FTC’s 2025 study warned that cloud‑provider investments risk “locking in the market dominance of large incumbent technology firms.” Senators Warren and Wyden have sent letters probing both the Google-Anthropic and Microsoft-OpenAI relationships, and the DOJ and FTC launched a joint inquiry on competitor collaborations in February.

For developers building on Claude, the investment is unambiguously positive in the medium term. The 10‑gigawatt compute expansion directly addresses the capacity constraints that have caused API rate limits, usage caps, and the.¹ Here’s what to expect: