Anthropic adapts its AI model to AMD GPU in just 3 days, challenging NVIDIA CUDA's competitive edge

Glonhui, November 12th — According to ITHome, the leading advantage of Nvidia's CUDA ecosystem may be facing increasing challenges. A slide from Anthropic's IPO roadshow materials shows that as high-performance AI models gradually gain the ability to autonomously adapt to different computing platforms, GPUs are becoming more like interchangeable general-purpose computing resources. Data shows that Anthropic's Claude Mythos Preview model figured out, in just about three days, how to run another model on AMD MI355X GPU. In contrast, traditionally, model adaptation for inference tasks on a new computing platform typically requires 6 to 12 months of work. This means that Anthropic can now treat GPUs from different manufacturers as relatively universal computing resources, and the technical barriers to switching between platforms have been significantly lowered. Consequently, there are now opinions that Nvidia's long-held proud CUDA ecosystem moat may, in fact, have already lost its original significance.