Anthropic is expanding its partnership with Amazon Web Services (AWS), deepening their work together to develop and deploy advanced AI systems. This expanded partnership includes a new $4 billion investment from Amazon and establishes AWS as Anthropic’s primary cloud and training partner. This will bring Amazon’s total investment in Anthropic to $8 billion, while maintaining their position as a minority investor.
Anthropic is working closely with Annapurna Labs at AWS on the development and optimization of future generations of Trainium accelerators, advancing the capabilities of specialized machine learning hardware. Through deep technical collaboration, we’re writing low-level kernels that allow us to directly interface with the Trainium silicon, and contributing to the AWS Neuron software stack to strengthen Trainium. Engineers work closely with Annapurna’s chip design team to extract maximum computational efficiency from the hardware, which they plan to leverage to train the most advanced foundation models.
This close hardware-software development approach, combined with the strong price-performance and massive scalability of Trainium platforms, enables us to optimize every aspect of model training from the silicon up through the full stack.
Claude as core infrastructure
Through Amazon Bedrock, Claude has become core infrastructure for tens of thousands of companies seeking reliable, practical AI solutions at scale. Pfizer uses the latest Claude models in Amazon Bedrock to accelerate research and delivery timelines for critical medicines, while saving tens of millions in operational costs. Intuit uses Claude in Amazon Bedrock to explain complex tax calculations for millions of users during tax season. Perplexity, an AI-powered search engine, delivers more accurate responses at twice the speed by using Claude in Amazon Bedrock, and the European Parliament leverages Claude to power ‘Archibot’, making 2.1 million official documents instantly searchable and easier to analyze in multiple languages while reducing research time by 80%.
Enabling secure, customizable AI solutions
Claude in Amazon Bedrock provides access to frontier intelligence within AWS, allowing customers to keep models and data in the same cloud environment. Building on this unified architecture, organizations can fine-tune Claude models in Amazon Bedrock, including Claude 3 Haiku, to tailor outputs to their specific requirements and boost accuracy for domain-specific tasks. Customer proprietary training data remains secure within AWS, preserving our commitment to safety and privacy.
By leveraging AWS’s robust security features and compliance certifications, organizations can confidently deploy AI solutions that meet stringent regulatory requirements. Government customers and industry partners can access Claude’s capabilities through Amazon Bedrock in AWS GovCloud (US). Additionally, government customers can access Claude through Amazon SageMaker in highly controlled environments like the AWS Secret and Top Secret Cloud Regions.
