If there’s anything in the AI space I would expect to forment a crash, it’s the TPU/trainium bullshit being pushed by UHNW forces. Nobody wants to even show up for ROCM, let alone a fully NDA’d solution.
Rumor says Hudson River Trading just ordered a bunch. So, the finance AI guys definitely. And they (AI finance - DeShaw, HRT, Citadel, G Research, XTX) have deployed about 15% of total GPU capacity, so not small fries.
the hyper scalars are buying huge AMD they figured out how to do inference on them with out great software bugs. but on an average nividia is beating them out ! google is on a different level though.
If there’s anything in the AI space I would expect to forment a crash, it’s the TPU/trainium bullshit being pushed by UHNW forces. Nobody wants to even show up for ROCM, let alone a fully NDA’d solution.
Rumor says Hudson River Trading just ordered a bunch. So, the finance AI guys definitely. And they (AI finance - DeShaw, HRT, Citadel, G Research, XTX) have deployed about 15% of total GPU capacity, so not small fries.
jane street too btw.
> Nobody wants to even show up for ROCM
I hear this over and over yet AMD's AI revenue dwarfs its consumer revenue.
Somebody is buying AMD cards for AI. I would love to know who.
the hyper scalars are buying huge AMD they figured out how to do inference on them with out great software bugs. but on an average nividia is beating them out ! google is on a different level though.
I mean, they work? MI355 are quite good for inference.