Microsoft's cooking up a new AI chip codenamed Athena¹⁷. Sources say it could drop as early as next month. Microsoft and Open AI have been testing this new baby for a while.

What’s going on here?

Microsoft could launch a new AI chip at its annual developer conference.

What does this mean?

This new chip, Athena, is meant for giant data centres and training those large language models we all know and love. Microsoft started working on Athena around 2019, years after Amazon and Google (right around the time they invested $1B in Open AI).

The game plan is to offer Athena through Azure cloud and reduce costs on all things AI. Their hope is to match Athena’s performance to Nvidia's H100 GPUs. Microsoft has also teamed up with AMD on their upcoming AI chip. Seems like a well-thought process to reduce reliance on its dear pal Nvidia.

Why should I care?

Owning the LLM hardware means Microsoft can accelerate its AI capabilities and push Azure ahead harder. This two-pronged approach - partnerships and proprietary tech - is textbook Microsoft strategy. Open AI wants to make custom hardware too⁷. But Microsoft's move shows building competitive AI chips takes years.

And Nvidia needs to watch out. Despite hitting a trillion-dollar valuation, interest in alternative AI chips is rising. With big tech making their own, and startups looking at Intel or AMD, Nvidia's future growth might take a hit.