News Amazon's $8 billion Anthropic investment rumors suggest it would rather sell AI infrastructure than compete with ChatGPT and Gemini

Which makes a lot of sense. Google has decades of experience in search that it can use for Gemini. OpenAI is the segment leader. Amazon is the leader in infrastructure as a service, but their software offerings have been historically weak. A good example of this is their recent phasing out of Workdocs and adoption of Microsoft 365 internally. Sends a message to your clients when your software offering is not good enough for your own teams.

They are better off selling infrastructure.
 
Anthropic is the most powerful and most popular model available on AWS Bedrock

For AWS customers who want or require their models running privately in their own environment, or GovCloud customers, Anthropic is their best option. OpenAI and Google aren't even an option for them to use at all because OpenAI doesn't provide their models to others to host to anyone but Microsoft, and Google doesn't provide theirs to anyone at all (so you would need to move your whole workload to Google).

Your only options become Anthropic, DeepSeek (Chinese), Meta, IBM, and a few others that no serious company will use.

As such... AWS investing in Anthropic just makes sense as it's a partner their customers will continue to use, and meets their security and compliance needs (which they can't do with OpenAI, unless they access it with Azure). It's a win/win/win scenario.