![](/static/253f0d9b/assets/icons/icon-96x96.png)
![](https://fry.gs/pictrs/image/c6832070-8625-4688-b9e5-5d519541e092.png)
You’re not wrong that GPU and AI silicon design are tightly coupled, but my point was that both of the GPU manufacturers are dedicating hardware to AI/ML in their consumer products. Nvidia has the tensor cores in its GPUs that it justifies to consumers with DLSS and RT but we’re clearly designed for AI/ML use cases when they presented them with Turing. AMD has the XDNA AI Engine that it is putting its APUs separate from its RDNA GPUs
Nvidia does not have a strong history of open sourcing things, to say the least. That last bit sounds like pure hopium