Google’s $40 Billion Anthropic Deal Signals Massive AI Power Shift

Google (Alphabet) is planning to spend up to 40 billion dollars on AI start-up Anthropic. This confirms Google as its strategic primary investor and adds it to an impressive list of investors that includes, to name a few, Amazon and Microsoft.

The Google - Anthropic partnership is undoubtedly targeted at market leader OpenAI, whose ChatGPT is approaching a billion users. And while Google is moving into the B2C market with Gemini, Anthropic is already the leader in B2B.

Structure of the Investment

Based on information from Anthropic, Google is initially paying 10 billion dollars in cash. Anthropic is valued at 350 billion dollars, the same as the February funding round. An additional 30 billion dollars is to be paid under certain performance conditions.

On top of the investment, Google is also significantly increasing its infrastructure services. As part of Google Cloud, it will provide Anthropic with 5 gigawatts of computing power over the coming five years, with the possibility of expansion.

A Remarkable Cap Table

Google has been an investor in Anthropic for some time now and is therefore part of an exclusive group of investors. Google is joined by Amazon and Microsoft as the most notable investors in the field of AI, and this week, Amazon announced a further 5 billion investment in Anthropic. In exchange for its investment, Anthropic is set to invest up to 100 billion dollars in Amazon’s cloud computing services.

This group is interesting as Google is both a competitor and an investor in Anthropic. Google has both its own model (Gemini) and, as part of Google Cloud, provides key infrastructure to Anthropic - and in particular access to its proprietary so-called Tensor Processing Units (TPUs) chips.

Strategic Background

The funding takes place against the backdrop of the competition for computing power for AI applications. Anthropic recently experienced a shortage of capacity of its AI assistant Claude and, in response, signed several infrastructure deals, including one with cloud computing company CoreWeave and the extended agreement with Google and chip maker Broadcom for TPU computer capacity from 2027 onwards.

In February, Anthropic was last valued at 350 billion dollars. Sources have now indicated that investors are willing to pay 800 billion dollars or more for Anthropic. There are also suggestions that Anthropic may be eyeing an IPO in October.

Newest Model, Limited Access

As well as the funding announcements, Anthropic has also just released its most advanced model to date, Mythos, to a select group of partners. The model reportedly has several applications in cybersecurity, so Anthropic is only making it available to the public on a limited basis to assess security risks with several organisations. There have been reports, however, that the model has already been leaked.

#GoogleAnthropic #AnthropicAI #AIInvestment #CloudComputing #OpenAIRivalry