Microsoft Apparently Wants to “Break” Nvidia’s Moat, Making CUDA Available to AMD AI Chips
SMRTR summary
A Microsoft employee claims the tech giant is secretly building software tools to break NVIDIA's iron grip on artificial intelligence computing. The alleged toolkit would translate code from NVIDIA's proprietary CUDA platform to run on cheaper AMD chips instead.
According to a transcript circulating online from what appears to be an industry expert forum, the anonymous Microsoft worker described growing pressure to slash AI costs over recent months. "We built some toolkits to help convert like CUDA models to ROCm so you could use it on an AMD, like a 300X," the person said.
The move would strike directly at NVIDIA's "golden cage" - its CUDA software that keeps AI developers locked into expensive NVIDIA hardware. For years, this ecosystem has maintained the company's dominance even as competitors offer cheaper alternatives.
Microsoft's alleged strategy mirrors a broader industry revolt. Oracle recently announced plans to deploy 50,000 AMD chips, while OpenAI signed a multi-billion dollar agreement with AMD that could give the AI lab a 10% stake in the chipmaker.
The shift reflects mounting pressure as AI inference costs balloon across millions of users, forcing tech giants to seek alternatives to NVIDIA's premium-priced silicon.
SMRTR provides this summary for quick context. The original article belongs to WinBuzzer.
Read the original article