Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Anthropic has reached a deal to secure access to 1mn Google Cloud chips to train and run its artificial intelligence models, increasing its ties to one of its largest investors.
Google, which has invested more than $3bn in Anthropic, will bring more than a gigawatt of AI computing capacity online for the start-up next year using its custom chips known as Tensor Processing Units, or TPUs.
Anthropic said the deal was worth tens of billions of dollars, but would not give a specific estimate.
“Anthropic and Google have a long-standing partnership, and this latest expansion will help us continue to grow the compute we need to define the frontier of AI,” said Krishna Rao, Anthropic’s chief financial officer.
“This expanded capacity ensures we can meet our exponentially growing demand while keeping our models at the cutting edge of the industry,” he added.
The agreement follows a flurry of deals by Anthropic’s chief rival OpenAI to secure chips and computing capacity from Nvidia, AMD, Broadcom, Oracle and Google, estimated to be worth about $1.5tn.
The circular arrangements between companies that act as suppliers, investors and customers of each other, combined with booming AI valuations, have added to concerns about a bubble in the sector.
AI model developers have accelerated their fundraising and dealmaking for fear of falling behind in the arms race to secure enough computing power to remain competitive and meet customer demand.
Anthropic raised $13bn during a funding round that closed in September, lifting its valuation to $183bn. It remains dwarfed by OpenAI’s $500bn valuation.
San Francisco-based Anthropic, which makes the Claude chatbot, uses three different chip platforms to train and run its AI systems: Amazon’s Trainium, Nvidia’s GPUs and Google’s TPUs.
The start-up said the diversified approach meant “we can continue advancing Claude’s capabilities while maintaining strong partnerships across the industry”.
The arrangement pits Amazon, Nvidia and Google against each other in the competition for huge contracts to supply computing power.
“Anthropic’s choice to significantly expand its usage of TPUs reflects the strong price-performance and efficiency its teams have seen with TPUs for several years,” Thomas Kurian, chief executive at Google Cloud, said in a statement.
Amazon is the start-up’s “primary” cloud provider and a large investor in the company. It has invested $8bn in Anthropic and is building a 2.2GW data centre cluster in New Carlisle, Indiana, to help train its AI models.
Earlier this year it weighed investing more in the start-up to deepen its relationship with the model developer, the Financial Times reported.
Anthropic said that it remained committed to its partnership with Amazon.