Microsoft Corp. and Nvidia Corp. plan to invest in Anthropic AI under a new tie-up that requires the Claude maker to use Microsoft Azure, in yet another high-profile AI deal amid fears of a bubble.
According to the Microsoft-Nvidia-Anthropic deal, Nvidia and Microsoft will invest $10 billion and $5 billion, respectively, in Anthropic which in turn will buy $30-billion worth of Azure compute capacity, according to a statement on Tuesday. Separately, Nvidia and Anthropic will establish “deep technology partnership” that includes collaboration on design and engineering.
“We are very excited to get additional capacity that we can use both to train our models to support Microsoft first party products and sell together,” Anthropic CEO Dario Amodei said in a YouTube video accompanying the announcement.
AI boom
The Microsoft-Nvidia-Anthropic deal underscores the insatiable appetite for compute required to power an AI world. It also ties Microsoft and Nvidia closer to a rival of OpenAI Inc.—in which both Satya Nadella and Jensen Huang are significant investors.
Earlier, OpenAI announced a $38-billion deal with Amazon.com Inc. that gave the ChatGPT maker access to Amazon Web Services. CEO Sam Altman has said that the world’s most valuable private firm is committed to spend $1.4 trillion to develop 30 GW of AI compute power.
AI bubble
To be sure, investors are increasingly uneasy that the AI boom has outrun its fundamentals. Some business leaders have noted that circular deals—where one partner props up another’s revenue—add to the bubble risk.
A few large investors have dumped some of their AI holdings, including Peter Thiel and SoftBank Group Corp., even as ‘Big Short’ investor Michael Burry emerges as the biggest critic of the AI deals.
Burry’s Scion Asset Management has bought put options on Nvidia and Palantir, positions that would benefit from declines in their share prices. Burry, who rose to fame for his 2008 bet against the US housing market, also posted a cryptic message on X: “sometimes, we see bubbles”.
