AI compute is becoming the industry’s most expensive assumption. Iren’s view that demand may never be fully satisfied captures the mood around neoclouds: every new model, agent, and AI product seems to create another reason to buy more GPUs, power, and data-center capacity.
That optimism is also a risk signal. If supply creates more demand, infrastructure companies can grow quickly. If usage, pricing, or model efficiency shifts, the same buildout can become a debt-heavy race where only the best-capitalized providers survive.
For Pagish readers, this is the physical layer underneath every model announcement. The next AI winners may be shaped as much by who controls reliable compute as by who writes the best model card.
Was this useful?
Help Pagish understand which AI stories are worth covering more deeply.
Tell Pagish if this story was useful.