Amazon has agreed to invest up to $4 billion in Anthropic, pairing funding for the AI startup with a closer role for Amazon Web Services in building and distributing its models. The arrangement links investment, cloud computing and customer access as both companies expand their work in generative AI.
Funding comes with a larger option
Amazon will initially invest $1.25 billion for a minority stake in Anthropic. It also has the option to increase its investment to a total of $4 billion. The companies did not disclose how Amazon valued Anthropic as part of the new deal.
The agreement adds to Anthropic’s funding history. The startup had raised $2.7 billion to date and was valued at about $5 billion in May this year, when it secured $450 million in a funding round. TechCrunch reported earlier this year that Anthropic planned to raise as much as $5 billion over the next two years.
Anthropic develops Claude, an AI-powered chatbot that analyzes text. The company launched its first consumer-facing premium subscription plan for Claude 2 earlier this month. It has also described plans for a “frontier model,” tentatively called “Claude-Next,” that its 2023 investor deck said would be 10 times more capable than today’s most powerful AI.
AWS becomes part of Anthropic’s model development
Anthropic will use AWS as a primary cloud provider for mission-critical workloads, including safety research and future foundation model development. It will also use Amazon’s Trainium and Inferentia chips to build, train and deploy future foundation models. Anthropic has been an AWS customer since 2021.
That arrangement gives Anthropic access to cloud computing and specialized chips for its research and model work. For Amazon, the partnership brings a prominent AI developer further into its cloud ecosystem, including the infrastructure used to develop and run models.
Anthropic has committed to providing AWS customers around the world with access to future generations of its foundation models through Amazon Bedrock. The company also plans to offer those customers early access to unique model customization and fine-tuning capabilities.
What the agreement offers AWS customers
Amazon Bedrock is AWS’s managed service for using foundation models to build generative AI applications. The source describes the service as providing secure access to leading foundation models. Anthropic’s commitment means its future models will be available through that service to AWS customers.
Amazon CEO Andy Jassy said customers were excited about Bedrock and Trainium, Amazon’s AI training chip, and that the collaboration should help them get more value from those capabilities. The deal therefore connects Anthropic’s model development with Amazon’s tools for businesses that want to build AI applications.
Anthropic said that training state-of-the-art models requires extensive computing resources and research programs. It described Amazon’s investment and supply of Trainium and Inferentia technology as support for advancing AI safety and research, as well as scaling adoption of Claude.
Resources for an ambitious AI push
The partnership arrives as Anthropic pursues more capable models. The startup cautioned that its planned frontier model would require a billion dollars in spending over the next 18 months. Amazon’s investment and cloud resources could support that effort, while the agreement creates a path for future Anthropic models to reach AWS customers.
Anthropic also has a range of other backers, including Google, Spark Capital, Salesforce, Sound Ventures, Menlo Ventures and Zoom. Amazon’s deal adds a strategic investor that is also a cloud provider and chip supplier, tying financial support to infrastructure and distribution.
For Amazon, the deal deepens its involvement in a fast-growing AI sector and supports its cloud offering. For Anthropic, it combines capital with access to computing capacity and a route to business customers. The companies’ arrangement makes those links central to how they plan to develop and deliver future AI models.