Anthropic, OpenAI’s close rival, has raised an additional $4 billion from Amazon, and has agreed to make Amazon Web Services (AWS), Amazon’s cloud computing division, the primary place it’ll train its flagship generative AI models.
Anthropic also says it’s working with Annapurna Labs, AWS’ chipmaking division, to develop future generations of Trainium accelerators, AWS’ custom-built chips for training AI models.
“Our engineers work closely with Annapurna’s chip design team to extract maximum computational efficiency from the hardware, which we plan to leverage to train our most advanced foundation models,” Anthropic wrote in a blog post. “Together with AWS, we’re laying the technological foundation — from silicon to software — that will power the next generation of AI research and development.”
In its own post, Amazon clarified that Anthropic will use Trainium to train its upcoming models and Inferentia, Amazon’s in-house chips to accelerate the running and serving of models, to deploy those models.
The new cash infusion from Amazon brings the tech giant’s total investment in Anthropic to $8 billion, but maintains Amazon’s position as a minority investor, Anthropic says. To date, Anthropic has raised $13.7 billion in venture capital, according to Crunchbase.
The Information reported earlier this month that Amazon was in talks to invest multiple billions in Anthropic, its first financial pledge in the company since a $4 billion deal struck last year. The new investment is reportedly structured similar to the last one — but with a twist. Amazon insisted that Anthropic use Amazon-developed silicon hosted on AWS to train its AI.
Anthropic is said to prefer Nvidia chips. But the money may have been too good to pass up. Around the start of the year, Anthropic reportedly projected it would burn through over $2.7 billion in 2024 as it trained and scaled up its AI products. Anthropic has for several months been discussing new funding at a $40 billion valuation, per The Information, and no doubt the pressure was on to clinch something soon.
Anthropic notes that its work with AWS has expanded over the years. Through Amazon Bedrock, AWS’ platform for hosting and fine-tuning models, Anthropic’s Claude family of models are being used by “tens of thousands” of companies, Anthropic says.
Recently, Anthropic teamed up with AWS and Palantir to provide U.S. intelligence and defense agencies access to Claude.
Beyond AWS, Amazon is said to be working with Anthropic to revamp the former’s consumer products. Reportedly, Amazon is set to replace the in-house models powering Alexa, its virtual assistant, with Anthropic’s after encountering technical challenges.
The collaboration — and investments — have attracted regulator scrutiny.
The FTC earlier this year sent a letter to Amazon, as well as Microsoft and Google, requiring the companies to explain the impacts their investments in startups such as Anthropic have on the competitive landscape of generative AI. Google has also invested in Anthropic, having poured $2 billion into the company late last October and taken a 10% stake, while Microsoft is a major OpenAI backer.
The U.K.’s competition regulator, the Competition and Markets Authority, which has also opened several inquiries into big tech tie-ups with AI upstarts, recently okayed Alphabet’s partnership and investment Anthropic after giving Amazon’s investment last year the green light.
Anthropic continues to maintain pace with the other frontier AI labs, including OpenAI, releasing new capabilities like Computer Use, which lets its current best model perform tasks on a PC autonomously. But the company has faced its share of setbacks, as well. It unexpectedly hiked the price of one of its models recently. And it’s seen the timetable slip for the release of its next-generation, top-of-the-line Claude model, 3.5 Opus.
