Amazon pumped $2.75bn into the start-up last year, completing a $4bn deal with the company.

AI start-up Anthropic is nearing a deal to raise $2bn in a funding round led by Lightspeed Venture Partners, a move that would place the three-year old company at a $60bn valuation, sources told the Wall Street Journal.

This comes after a $4bn injection deal from Amazon into the start-up in September 2023, which included a $1.25bn initial investment by Amazon Web Services (AWS) the same year, giving the e-commerce giant a minority stake in the business. The initial investment was followed with Amazon pumping $2.75bn into Anthropic in March 2024.

As part of the deal, Anthropic agreed to use AWS Trainium and Inferentia chips to build, train and deploy its future foundation models, while the two also announced that they will work together on improving the chips in the future.

Also in 2023, Google’s parent Alphabet invested $300m in Anthropic for a 10pc stake, which resulted in Anthropic making Google Cloud its “preferred cloud provider”.

While earlier last year, the UK’s Competition and Markets Authority which investigated an Alphabet-Anthropic partnership in the country, concluded that it did not qualify for an investigation, citing that Alphabet would not have “material influence” over the AI start-up as a result of the partnership.

Anthropic, the start-up behind the AI chatbot Claude, burst onto the industry at a time when businesses were beginning to jump on the generative AI (GenAI) train.

Created by several former OpenAI employees in 2021, the ‘public benefit corporation’ calls itself an “AI safety and research company”, even developing a framework for assessing different AI capabilities to respond to emerging risks.

The company is often viewed as an alternative to the GenAI giant OpenAI, a company embroiled in a number of controversies shifting towards becoming a for-profit corporation, breaking away from its previous non-profit company structure.

However, Anthropic has come under fire for claims of copyright infringement. Last year, a trio of authors took the start-up to court over claims that the company used pirated versions of various copyrighted material to train Claude.

In a filing with a US district court in California, the plaintiffs claim that the “largescale theft of copyrighted works” is a key component of Anthropic’s business model, adding that the company has “taken multiple steps to hide the full extent of its copyright theft”. Anthropic however, denies these claims.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.