The futures market taking shape around computing power could amplify an AI price crash as easily as it cushions one, Transformer's Conrad Quilty-Harper reported.
CME Group, Intercontinental Exchange and Architect Financial Technologies are all preparing contracts tied to GPU rental prices, pitched as a hedge for the lenders behind a buildout McKinsey estimates will require $7 trillion of data centers by 2030.
But a public price for compute also gives speculators something to short. Berkeley law professor Frank Partnoy told Transformer that "if you create the market, inevitably people will use it to bet on these prices."
Transformer points to the margin spirals that hit nickel and UK pension funds as the template for what happens when a hedge turns into a forced seller. Oxford's Rama Cont said the bigger danger is the debt behind the buildout.
The DeAI takeaway: Fast-moving markets could strengthen the case for decentralized computing networks that can quickly redistribute capacity from providers with spare hardware to value-conscious users who are comfortable being price-takers.