Receive free Artificial intelligence updates
We’ll send you a myFT Daily Digest email rounding up the latest Artificial intelligence news every morning.
Amazon plans to invest up to $4bn in artificial intelligence start-up Anthropic, as the big tech group steps up its rivalry with Microsoft, Google and Nvidia to persuade AI companies to use its technology.
The deal, announced on Monday, will see Amazon invest an initial $1.25bn for a minority stake in Anthropic. Both companies will later have the option to increase the total to $4bn, they said.
It is a bid to forge a close relationship with a prominent AI start-up akin to Microsoft’s alliance with OpenAI, the group behind ChatGPT. As part of the agreement, Anthropic will use Amazon’s cloud computing platform and its dedicated AI chips to create its models.
AI start-ups have become locked in an arms race to secure the costly chips and data centre resources necessary to build the latest AI systems, called large language models.
Anthropic was last valued at nearly $5bn in a funding round earlier this year, according to a person familiar with the terms. It is one of the top competitors to OpenAI raising huge sums this year, including Inflection AI, which raised $1.3bn from Microsoft and Nvidia, and Toronto-based Cohere, which raised $270mn from Nvidia and others.
The start-up’s new alliance with Amazon Web Services appeared to be a shift away from Google, which invested $300mn in Anthropic last year. It comes just seven months after Anthropic said it would train its models on Google’s chips and use its cloud.
For Amazon, the deal marks its latest push to capitalise on the excitement around generative AI, a technology capable of creating humanlike text and realistic images. Amazon is looking to position its Trainium and Inferentia chips as credible alternatives to Nvidia’s processors for the training and running of generative AI models.
Anthropic’s Claude chatbot, which rivals ChatGPT, is already among a range of AI products available on AWS’s Bedrock service, which allows customers to build generative AI applications in the cloud.
The new deal represents a “significant expansion of the partnership” with Anthropic, said Adam Selipsky, head of AWS. Anthropic would have access to “significant quantities” of Trainium chips to train future versions of its foundation models, he added.
Under a landmark partnership with OpenAI earlier this year, Microsoft invested billions in the Silicon Valley-based start-up and became its exclusive cloud provider. While Anthropic’s deal with Amazon has echoes of that agreement, it is not an exclusive arrangement.
Selipsky said Amazon’s strategy was “very different” to Microsoft’s, as it centred around “providing customers with the most choice and most security . . . Exclusivity is not the goal.”
Amazon will be the Anthropic’s “primary” cloud provider for “mission critical workloads”, the companies said.
Anthropic’s co-founder Dario Amodei insisted “nothing has changed” in its arrangement with Google, announced in February. At that time, Google said that it was Anthropic’s “preferred cloud provider”.
Chipmaker Nvidia has so far been the biggest winner from the frenzy of excitement about generative AI.
Jim Hare, a technology analyst at Gartner, said Amazon’s deal with Anthropic would allow AWS to demonstrate that Nvidia’s chips were not the only option for AI developers.
With Monday’s deal, Amazon could point to “a model provider that has cachet” using its chips, Hare said. It may help dispel the “perception” that Amazon had fallen behind on AI, he added.
Anthropic was founded when a group of OpenAI employees, led by Amodei, split from the company after differences over the group’s direction, following its $1bn investment from Microsoft in 2019.
Amodei said: “I feel very good about this (Amazon) deal and the provisions in it to make sure that things stay safe.”
The valuation at which the investment is being made, and the size of Amazon’s ownership stake, will be set at a later funding round. Two people familiar with the terms said it would leave Amazon with a far smaller stake than the 49 per cent ownership that Microsoft is said to have sought with its investment in OpenAI.