Here's the split in AI compute that not many are reading correctly.

Frontier training is concentrating harder every quarter, thousands of GPUs that have to sit in one place wired together. But training is only 30% of demand in 2026. The other 70% is inference, and running it on a hyperscaler means paying for infrastructure built for the hardest workload to do the easiest one.
On distributed networks that same inference could run 45-75% cheaper and for anyone sizing an AI infrastructure budget, that gap is the whole story.
Training centralizes by necessity. Inference fragments because paying AWS margins for a workload that doesn't need them stops making sense at scale.