Microsoft and AMD join forces to challenge Nvidia's AI grip

Nvidia owns the generative AI boom right now, but Microsoft might be looking to groom its biggest rival.

According to Bloomberg, Microsoft is teaming up with AMD to boost the chipmaker's AI capabilities, chipping in both technical support and engineering resources.

While Microsoft already runs tens of thousands of Nvidia GPUs across its Azure data centers, relying on a single supplier backed by a CUDA monopoly is getting pricey.

A secret chip and a swift denial

The report also claimed that AMD is helping Microsoft develop its own in-house AI chip codenamed Athena, which was first reported by The Information last month.

However, a Microsoft spokesperson flatly pushed back, denying that AMD is involved in Athena.

Denial aside, the motive behind backing AMD is crystal clear. Microsoft is reportedly working on a private enterprise version of ChatGPT that could cost ten times as much to run as the regular version. Cheaper chips mean cheaper servers, which directly protects Microsoft's bottom line.

Why it matters

For businesses looking to deploy enterprise AI, genuine hardware competition could eventually keep sky-high operating bills from getting out of hand.

For the industry, Nvidia’s real moat has always been software lock-in through CUDA. AMD has long struggled with AI developer adoption, and having Microsoft’s engineers directly optimize its silicon could finally give Nvidia a run for its money.

We think Microsoft is playing smart 4D chess here: buy Nvidia cards today, build custom chips for tomorrow, and prop up AMD to keep everyone honest. That said, closing the CUDA gap takes years, so Microsoft will remain dependent on Nvidia for the foreseeable future.