Natural language processing

AWS invests another $4 billion in Anthropic to push its own AI hardware and cloud

AWS invests another $4 billion in Anthropic to push its own AI hardware and cloud



summary
Summary

Amazon is doubling down on its investment in OpenAI competitor Anthropic, bringing its total investment to $8 billion. The partnership aims to solidify Amazon Web Services (AWS) as the go-to infrastructure for AI model training.

As part of the deal, Anthropic will use AWS Trainium and Inferentia chips to train and deploy its foundation models. The company’s engineers are working directly with AWS’s Annapurna Labs team to develop low-level kernels for hardware communication and optimize the AWS Neuron software stack.

AWS says that tens of thousands of companies already utilize Anthropic’s Claude AI models through the Bedrock AI cloud service. Customers can also use Amazon Bedrock to train Claude models, like Claude 3 Haiku, with their own data.

More than just an AI investment

This new investment, building on an initial $4 billion investment from September 2023, underscores Amazon’s strategic focus on AWS. While the exact details remain undisclosed, the deal likely involves a combination of cash and AWS computing credits, mirroring the previous agreement.

Ad

This approach, similar to Microsoft’s investment in OpenAI, isn’t solely about AI advancement; it’s a calculated move to drive AWS growth.

Because Anthropic uses AWS infrastructure, a significant portion of the investment flows back to Amazon, increasing its cloud revenue. This is critical for Amazon, as investors are closely watching the growth of its cloud business. Major cloud providers like Google, Microsoft, and Amazon are heavily invested in the AI boom precisely because of this symbiotic relationship.

Think of it like this: an electricity grid operator invests in a light bulb manufacturer to develop power-hungry bulbs. The more power the bulbs consume, the more revenue the utility generates. Similarly, cloud providers benefit from compute-intensive AI models. As AI applications require more computing power, the demand for cloud resources increases, which directly benefits AWS.

Anthropic expands into defense with AWS and Palantir

The two companies also recently announced a collaboration to bring their AI models to the defense sector. A new partnership gives U.S. intelligence agencies access to Anthropic’s Claude AI models through Palantir’s AI platform, hosted in a secure AWS environment designed for sensitive government data.

Anthropic’s terms of service explicitly allow use for intelligence analysis, detection of covert operations, and early warning of military activity. Meta and OpenAI are also pursuing collaborations in the defense sector.

Recommendation

The UK Competition and Markets Authority has greenlit the $4 billion investment, citing Anthropic’s limited UK market presence. However, the US Federal Trade Commission is reviewing the deal as part of a broader examination of AI partnerships involving major tech companies, including the Microsoft-OpenAI collaboration and Google’s earlier $2 billion investment in Anthropic.

AWS invests another $4 billion in Anthropic to push its own AI hardware and cloud

Source link