Startup emerges to help Wall Street price AI compute costs

Silicon Data has raised $30 million in Series A funding to launch the first compute futures trading on the CME on October 5, 2026, pending regulatory approval. The startup aims to establish a reference price for GPU rentals and an index for Wall Street futures. Despite regional halts on data centers in Texas and New York, Silicon Data research head Steve Hou asserts that data shows the AI buildout remains robust.

Silicon Data Series A funding

• ▪Silicon Data closed a $30 million Series A funding round to develop solutions for pricing and hedging artificial intelligence compute costs.
• ▪Silicon Data emerged to address the lack of a straightforward method for financial firms to price compute or hedge their exposure to changing compute costs.

GPU compute pricing benchmarks

• ▪Silicon Data aims to develop a benchmark index that Wall Street futures contracts can settle against.
• ▪Silicon Data aims to establish a standardized reference price for graphics processing unit (GPU) rentals.

Compute futures hedging instruments

• ▪Silicon Data plans to introduce compute futures trading to allow companies to hedge their financial exposure to fluctuating artificial intelligence compute costs.
• ▪Compute has become the single largest expense for developers of artificial intelligence products, driven by hundreds of billions of dollars spent annually on data centers and graphics processing units.

CME futures launch October

• ▪CME Group and Silicon Data partnered to launch the scheduled October 5, 2026 compute futures trading to unlock market pricing.
• ▪Silicon Data plans to launch its compute futures trading on the CME on October 5, 2026, subject to regulatory approval.

AI infrastructure buildout health

• ▪The artificial intelligence infrastructure sector faced headwinds in mid-2026 as Texas halted new data centers for audits and New York State suspended construction of all new data centers.
• ▪Silicon Data head of research Steve Hou stated that data shows the artificial intelligence buildout remains strong, challenging negative reports of depreciating chips and stalled data centers.