We know OpenAI and Azure was struggling to get enough GPU. That was implied not just by their words but also by action. And considering these two companies are most aggressive and making most money out of this AI. If GPUs are available they would have been able to buy it.
Build their own? It's what Microsoft, Google, and AWS are doing.
>Nvidia's sales are only limited by the number of wafers they can get from TSMC.
No, they're limited by the cost per operation vs. Facebook building their own. The cloud providers have already decided it's cheaper to do it themselves. Sure they'll keep buying GPUs for general public consumption but that may eventually end too.
There certainly is a lot of value in the CUDA ecosystem, today. The problem is that when all the big companies are buying up hundreds of thousands of GPUs, that doesn't leave much for anyone else.
Sane business people will look to decentralize their compute over time and not be reliant on a single provider. AMD will be able to take advantage of that and they've already stated that is their focus going forward.
ROCm/HIP are getting incrementally better, MI300x have 192GB and benchmarks are looking good, the only problem is that nobody has access to the higher end hardware from AMD today. That's why I'll have MI300x, for rent, soon.
That's a big issue in AMD land imho. Everyone can pickup a 200$ GPU (talking about the RTX 3050) which will behave like a scaled down A100 and get started playing around with CUDA. You can't really do that with AMD GPUs, their cheapest officially supported GPU is the 7900 XTX and that has a different architecture than the data center ones.
Nvidia has a vested interest in FB being beholden to their chips, so much so that it's worth giving them a discount to ensure it happens, and human nature being human nature a face saving discount has to be offered.
