Meet the startup helping Wall Street put a price on AI compute
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Silicon Data closed a $30 million Series A and aims to become the reference price for GPU rental — a benchmark comparable to Brent crude or the S&P 500, but for AI chip access.
- Silicon Data plans to launch compute futures trading on the CME on October 5th, pending regulatory approval, giving firms a formal vehicle to hedge GPU price swings.
- Compute has become the single biggest cost for anyone building AI products, with hundreds of billions of dollars a year flowing into data centers and GPUs, yet no standardized pricing mechanism currently exists.
- Silicon Data head of research Steve Hou, appearing on TechCrunch's Equity podcast alongside Rebecca Bellan, argued that the underlying data tells a healthier story than recent headlines about depreciating chips and stalled data centers.
- The startup directly addresses a gap in the market: firms building AI products have no straightforward way to price compute exposure or hedge when GPU rental rates shift.
Why it matters: Compute — the single biggest cost for AI builders — currently has no benchmark, leaving the hundreds of billions flowing into data centers and GPUs each year effectively unhedged. If Silicon Data's CME futures clear regulatory approval and debut October 5, it would give AI infrastructure investors and their customers the first formal mechanism to hedge chip price exposure, turning an opaque operational expense into a tradable commodity.
Ask SkimNews
