Amazon.com Inc. says it is investing another $2.75 billion in Anthropic, completing a deal made last year to back the artificial intelligence startup and expand a partnership between the companies.
The infusion brings Amazon’s total investment in Anthropic, a popular builder of artificial intelligence tools capable of generating text and analytics, to $4 billion, following an earlier investment announced in September. As part of that agreement, Amazon had the right to contribute the additional funds in the form of a convertible note, provided it did so by the end of March.
As part of the deal, Anthropic also agreed to use Amazon Web Services’ data centers to power some of its operations and to use Amazon’s custom computer chips. San Francisco-based Anthropic has also committed to using Google chips from Alphabet Inc., another close partner.
Anthropic has ties to several large tech companies, including Google, which last May joined a $450 million funding round led by Spark Capital. Google and Amazon Web Services are both cloud computing partners of Anthropic.
Anthropic was founded in 2021 by former OpenAI employees, including Daniela Amodei and her brother Dario, who serves as CEO. The company has since become one of OpenAI’s most formidable competitors, raising billions in funding. Most of its clients are businesses, from search engine DuckDuckGo to travel guide publisher Lonely Planet.
The company, which offers a chatbot called Claude, has emphasized the importance of developing artificial intelligence in a safe and responsible way. In early March, it introduced new chatbot software that it says will be better at carrying out complicated instructions and less likely to make things up.
Chatbots that can mimic human conversation have become a growing target of Silicon Valley companies, with rapid technological advances fueling an investment frenzy. Chatbots themselves are nothing new at all. But the technology that powers Claude and his competitors’ bots is a more powerful tool known as a large language model, which is trained on huge swaths of the Internet to generate text, such as an answer to a question or a poem. Such tools are an application of generative artificial intelligence, systems that consider input such as a text message and use it to produce new content.