one year on
Amazon completes $4B Anthropic investment, final $2.75B tranche closed
The e-commerce giant finalizes its $4 billion investment in Anthropic, locking in AWS as Anthropic's primary cloud provider and tying the startup's future model training to Amazon's Trainium and Inferentia chips.
Amazon today announced it has closed the final $2.75 billion tranche of its investment in Anthropic, bringing the total to $4 billion. The deal, first disclosed last September with an initial $1.25 billion, cements a strategic collaboration that makes AWS Anthropic’s primary cloud provider for mission-critical workloads, including safety research and future foundation model development.
As part of the agreement, Anthropic will use AWS Trainium and Inferentia chips to build, train, and deploy its future models, and has committed to providing AWS customers with ongoing access to its next-generation models via Amazon Bedrock. The Claude 3 family, already available on Bedrock, includes the Opus model which Amazon claims outperforms GPT-4 on reasoning, math, and coding benchmarks.
Amazon and Anthropic describe the deal as a strategic collaboration. “Generative AI is poised to be the most transformational technology of our time,” said Dr. Swami Sivasubramanian, vice president of Data and AI at AWS, in a statement.
The record
Said the strategic collaboration with Anthropic will further improve customer experiences and unlock opportunities for responsible innovation with generative AI.
One year later — open only if you can handle spoilers
Within two years, Anthropic's reliance on AWS chips became a point of tension as the startup explored multi-cloud strategies to negotiate better terms. The deal was widely seen as a template for later hyperscaler-AI startup investments, though critics noted that such arrangements often blur the line between partnership and control.