Last time I talked to our Azure reps, the technical folks seemed to be saying (or rather wanting to without explicitly saying so) that basically all present an

Last time I talked to our Azure reps, the technical folks seemed to be saying (or rather wanting to without explicitly saying so) that basically all present and near term GPU capacity (of recent gen) is either being consumed or will be consumed by OpenAI. Or in other words there is literally not enough GPUs on the planet (reasonably available) for MS to cover both OpenAI and Azure GPU loads without affecting the other. We basically couldn’t get them to commit to any price or delivery schedule if it involved the same hardware GPT needs. We are a $110B+ company btw and a strategic partner of MS, not some random shop. They’re not supposed to just say “no.”