Beginning of 2024, my employer was planning to buy GroqRacks, I knew someone from the CTO board and we had a chat about it. I was sceptical and din't hold back my opinion:
lock-in on niche technology, risk if the specific HW could serve upcoming/larger models, new model architecture implementations require proprietary knowledge and software, and we would need GPU clusters for training, anyway.
I don't know if that influenced the decision, but the project was stopped, in the end. I'm glad we didn't buy GroqHardware.
These IP games may work against competition elsewhere but they are largely impotent inside China.
While western economies are infighting, China is rapidly building out cheap renewable energy capacity.
The only missing piece China needs for global domination of energy intensive AI is a competitive GPU.
https://www.technologyreview.com/2025/07/10/1119941/china-en...