Amazon is putting more money and more infrastructure behind Anthropic. The AI company has raised an additional $4 billion from Amazon and agreed to train its flagship generative AI models primarily on Amazon Web Services, making AWS a central part of its next stage of model development.
The agreement is about more than capital. It links Anthropic’s future AI work to Amazon’s cloud platform, its Trainium accelerators, and its Inferentia chips for running models after training.
A bigger Amazon investment in Anthropic
The new funding brings Amazon’s total investment in Anthropic to $8 billion. Anthropic said Amazon remains a minority investor, even as the relationship between the two companies becomes more operationally important.
Anthropic has raised $13.7 billion in venture capital to date, according to Crunchbase. The company has been one of the major challengers to OpenAI, and its Claude family of models has become a key part of its commercial strategy.
The latest investment follows Amazon’s earlier $4 billion deal with Anthropic last year. That earlier deal was split into a $1.25 billion first installment last September and a $2.75 billion extension in March.
The Information reported earlier this month that Amazon was in talks to invest billions more in Anthropic. The new investment is reportedly structured similarly to the earlier one, but with one notable condition: Amazon wanted Anthropic to use Amazon-developed silicon hosted on AWS to train its AI.
Why AWS matters in the deal
Under the expanded partnership, Anthropic has agreed to train its flagship generative AI models primarily on AWS. That gives Amazon a deeper role in the infrastructure behind Anthropic’s most advanced systems.
Anthropic is also working with Annapurna Labs, AWS’ chipmaking division, on future generations of Trainium accelerators. Trainium is Amazon’s custom-built chip line for training AI models, and Amazon said Anthropic will use Trainium, including Trainium2, to train upcoming models.
Amazon also said Anthropic will use Inferentia, Amazon’s in-house chip for accelerating model running and serving, to deploy those models. In practical terms, the partnership spans both major phases of AI infrastructure: training models and serving them to users.
Anthropic framed the chip collaboration as a way to get more performance from Amazon’s hardware. In its blog post, the company said its engineers work closely with Annapurna’s chip design team to improve computational efficiency, with the goal of using that work for its most advanced foundation models.
Claude is becoming more tied to Amazon’s AI stack
Anthropic’s co-founder and CEO Dario Amodei said the company’s work with AWS has expanded substantially over the past year. Through Amazon Bedrock, AWS’ platform for hosting and fine-tuning generative models, Anthropic’s Claude family of models is being used by “tens of thousands” of customers.
That commercial relationship is now moving into a deeper infrastructure relationship. Amazon said AWS customers will get early access to the ability to fine-tune new Claude models on their data.
Anthropic has also recently 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 Amazon consumer products, and the tech giant is reportedly set to replace the in-house models powering Alexa with Anthropic’s after technical challenges.
For Amazon, Anthropic gives AWS a major AI partner at a time when cloud providers are competing to host and power frontier AI systems. For Anthropic, Amazon offers money, cloud distribution, and access to custom AI chips at a moment when training and scaling AI products is expensive.
The funding comes with pressure and scrutiny
The investment arrives as Anthropic faces heavy spending demands. Early this year, Anthropic reportedly projected it would burn through more than $2.7 billion in 2024 as it trained and scaled up its AI products.
The company has also been discussing raising new funding at a $40 billion valuation for several months, per The Information. That context helps explain why a large Amazon commitment would be significant, even if Anthropic is said to prefer Nvidia chips.
The growing collaboration between large tech companies and AI startups has drawn regulatory attention. The FTC earlier this year sent a letter to Amazon, Microsoft and Google requiring them to explain how investments in startups such as Anthropic affect competition in generative AI.
Google has also invested in Anthropic, pouring $2 billion into the company late last October for a 10% stake, while Microsoft is a major OpenAI backer. The U.K.’s Competition and Markets Authority has opened several inquiries into big tech tie-ups with AI firms, though it recently okayed Alphabet’s partnership and investment in Anthropic and had already given the green light to Amazon’s deal last year.
Anthropic keeps expanding while managing setbacks
Anthropic has continued to release new functionality while competing with other frontier AI labs. One recent feature, Computer Use, lets Claude 3.5 Sonnet perform tasks on a PC autonomously.
The company has also faced challenges. It unexpectedly hiked the price of one of its models this month and delayed the launch of Claude 3.5 Opus, its long-awaited, next-gen model.
To boost revenue, Anthropic has shifted some focus toward new tools and subscription plans. These include a desktop client, enterprise and “team” tiers, and mobile apps. It has also opened offices in Europe and made high-profile hires, including Instagram co-founder Mike Krieger, OpenAI co-founder Durk Kingma, and ex-OpenAI safety researcher Jan Leike.
Anthropic was co-launched in 2021 by Amodei, formerly VP of research at OpenAI, who reportedly split with the firm after disagreements over the roadmap. Amodei brought a number of ex-OpenAI employees to start Anthropic, including OpenAI’s former policy lead, Jack Clark.
The company often tries to position itself as more safety-focused than OpenAI. With Amazon’s latest $4 billion investment, Anthropic now has deeper backing for its AI ambitions, while AWS gains a closer connection to one of the most prominent companies building frontier generative AI models.