Amazon deepens AI partnership with Anthropic, commits additional $4 billion investment

Amazon has strengthened its partnership with AI research firm Anthropic, committing an additional $4 billion investment. Under the expanded collaboration Anthropic has named AWS its primary training partner and will utilize AWS Trainium and Inferentia chips to deploy its largest AI foundation models. This builds on last year’s agreement, which established AWS as Anthropic’s primary cloud provider. Anthropic was founded in 2021 by former OpenAI researchers including Dario Amodei and Paul Christiano, with a focus on developing safe and ethical artificial intelligence systems through their "constitutional AI" approach.

Anthropic’s Claude 3.5 models, including Haiku and Sonnet, are already available on AWS’s Amazon Bedrock platform facilitating enterprise users’ access cutting-edge AI capabilities. The collaboration now includes early fine-tuning access for AWS customers, allowing them to customize Claude with proprietary data.

Anthropic joins the competitive AI arms race along with OpenAI, Meta, and Google. OpenAI’s ChatGPT, hosted on Microsoft’s Azure cloud, remains a leader in generative AI, while Meta recently released its Llama models as open-source alternatives.

AWS CEO Matt Garman highlighted the rapid adoption of Anthropic’s models and praised the partnership’s potential to push the boundaries of generative AI. Meanwhile, Anthropic CEO Dario Amodei credited Amazon with enabling the scaling of its Claude family, which powers applications in translation, drug discovery, and engineering design.