Anthropic Discusses New Financing with Valuation Expected to Surpass $900 Billion

Bloomberg, citing people with knowledge of the matter, reports that Anthropic Corporation has begun weighing a new round of financing that would value the company at over $900 billion, potentially surpassing long-time competitor OpenAI to become the world's most valuable AI unicorn.

Anthropic is considering proposals put forth by investors, and if successful, its valuation would more than double from the existing level. These discussions are still in an extremely early stage, and the company has not accepted any offers.

It has been previously reported that Anthropic has rejected multiple proactive financing proposals from investors that valued the company at $800 billion or above.

This comes as Anthropic accelerates its fundraising efforts on the heels of its AI software's strong breakthroughs. It was previously reported that Anthropic is considering launching an initial public offering (IPO) as early as October this year, with the company actively seeking more infrastructure to meet the growing demand for its products.

Google recently pledged to invest $10 billion in Anthropic at a valuation of $350 billion, matching the valuation of its financing round in February this year. Google plans to add up to an additional $30 billion investment in Anthropic after it achieves certain performance targets. Amazon also invested $5 billion in Anthropic at a $350 billion valuation, with plans to inject another $20 billion over time. It is currently unclear whether these two companies will participate in this new round of financing.

Anthropic, founded in 2021 by a group of former OpenAI employees, has since joined the top ranks of the AI field. The company has developed a suite of AI tools aimed at disrupting the way businesses handle various tasks from programming to cybersecurity.

In early April this year, Anthropic Corporation released a new model called Mythos, which is said to detect and exploit vulnerabilities in all kinds of key software. Anthropic deems this model to be too risky for public release and currently only allows a handful of companies to start testing on their own systems. However, according to media reports, a small number of unauthorized users have accessed this model.