Laurent -

Amazon takes $4bn minority stake in AI safety research startup Anthropic

Amazon Web Services (AWS) will build on its existing ties to the firm to help it scale its enterprise-grade large language model (LLM) Claude

Amazon Web Services (AWS) is taking a minority stake worth $4bn in artificial intelligence (AI) safety and research startup Anthropic, and is set to become the firm’s primary cloud provider for mission-critical workloads.

Anthropic’s flagship offering is a generative AI system known as Claude that is aimed at enterprises, with business users encouraged to use their own data to train the system rather than rely on public datasets. This means enterprises can use their own proprietary data when training the large language model (LLM).

In a statement, announcing its work with AWS, Anthropic said the partnership is part of a “broader collaboration” that is geared towards developing the “most reliable and high-performing foundation models” in the technology industry.

The pair also said they want to ensure their foundation models are the safest and are put to use in a responsible way.

“AWS will become Anthropic’s primary cloud provider for mission-critical workloads,” the statement said, with the firm stating its intention to use the AWS Tranium second-generation machine learning accelerator and the AWS Inferentia chip to run its deep learning inference applications.

“Together, we’ll combine our respective expertise to collaborate on the development of future Trainium and Inferentia technology,” the statement continued.

The company has allowed AWS customers to build generative AI offerings of their own for some time using Claude with the help of its fully managed Amazon Bedrock offering, which makes foundation models from Amazon and startups like Anthropic available through an application programming interface (API).

This work will be expanding under the partnership, Anthropic confirmed: “Based on significant AWS customer demand for Claude, we’re also expanding our support of Amazon Bedrock. This will include secure model customisation and fine-tuning on the service to enable enterprises to optimise Claude’s performance with their expert knowledge, while limiting the potential for harmful outcomes.

“Amazon developers and engineers, will be able to build on top of our state-of-the-art models via Amazon Bedrock. This will enable them to incorporate generative AI capabilities into their work, enhance existing applications, and create net-new customer experiences across Amazon’s business.”

The past few months have seen Anthropic attract investment from a wide range of companies, including SAP and South Korea-based telco SK Telecom. The latter revealed in August 2023 that it would be ploughing $100m into the firm.

Several months before that, it was also announced that Anthropic had raised $450m in series C funding from investment firm, Spark Capital.

Closing out its statement, Anthropic said that, with Amazon’s financial backing and access to its Trainium and Inferentia technology, it will be better equipped to “continue advancing the frontier of AI safety and research.”

It added: “We look forward to working closely with Amazon to responsibly scale adoption of Claude and deliver safe AI cloud technologies to organisations around the world.”

Read more about cloud and generative AI

Read more on Infrastructure-as-a-Service (IaaS)

Data Center
Data Management