Alphabet commits up to $40 billion to Anthropic

Alphabet said on April 24 that Google would invest up to $40 billion in Anthropic, deepening one of the biggest financing relationships in artificial intelligence. Anthropic said the structure includes $10 billion in cash now and as much as $30 billion more if the startup meets performance targets. The deal links a company building the Claude model family with a company that also sells competing AI models under the Gemini brand. It also extends a partnership that already spanned equity investment, cloud distribution and custom chip infrastructure. ### How is the $40 billion actually structured? Anthropic said on April 24 that Google is committing $10 billion immediately, with another $30 billion available if the company hits performance milestones. Bloomberg reported that the upfront cash comes at a $350 billion valuation, while CNBC said the companies confirmed the broader up-to-$40-billion commitment on Friday. (cnbc.com) The milestone structure matters because the headline number is not all funded at once. CNBC reported that the remaining $30 billion is contingent on specified targets, and Bloomberg described the added money as tied to performance targets set by Anthropic and Google. ### Why is Google funding a company that also competes with Gemini? (cnbc.com) Google and Anthropic have been partners since 2023, when Google first invested $300 million for about a 10% stake, according to CNBC. CNBC also reported that Google later added another $2 billion and that, before the new commitment, its investment in Anthropic exceeded $3 billion. (cnbc.com) Google also distributes Anthropic’s Claude models through its cloud business, CNBC reported, even as Gemini competes with Claude in the market for AI models and services. That arrangement gives Google exposure to Anthropic’s growth as both a shareholder and an infrastructure provider. ### What does Anthropic get beyond the cash? (cnbc.com) Anthropic said the agreement will support a significant expansion of its computing capacity. CNBC reported that, earlier in April, Anthropic secured 5 gigawatts of computing capacity through an announcement with Google and Broadcom, with that capacity set to start coming online next year. (cnbc.com) Compute has become a central constraint for large AI companies because training and serving frontier models require enormous amounts of chips, power and data-center capacity. CNBC said Anthropic has been working to address what it called “inevitable strain” on infrastructure from rising enterprise, developer and consumer demand for Claude. (cnbc.com) ### How does this compare with other Anthropic backers? Amazon struck a similar deal with Anthropic just weeks earlier, CNBC reported. Under that agreement, Amazon invested $5 billion and agreed to invest up to $20 billion more tied to commercial milestones, CNBC said. The result is that Anthropic has lined up backing from multiple large technology companies at the same time it competes with them in AI services. (cnbc.com) Bloomberg described Google and Anthropic as both partners and rivals in the race to build artificial intelligence. ### What comes next in the arrangement? The next concrete milestone is infrastructure deployment. (cnbc.com) CNBC reported that the 5 gigawatts of computing capacity announced earlier this month will start to come online in 2027, and Anthropic said the new agreement is meant to expand computing capacity further. The remaining $30 billion from Google depends on whether Anthropic meets the performance targets tied to the deal. (bloomberg.com)