Ex-Citadel Quant Builds GPU Futures Platform to Hedge AI Demand Risk

As CoreWeave shares fall and drag much of the AI investment complex with them, a former Citadel Securities quantitative researcher is pitching a new class of financial instrument designed to let investors trade against a collapse in AI hardware demand.

Jasper Zhang, who completed a mathematics doctorate at UC Berkeley in two years before joining Citadel Securities to build trading models, now leads Hyperbolic, a 50-person fintech startup developing a spot marketplace for tradable GPU computing power.

The platform, built around software called Forge, allows investors to buy and sell GPU clusters — the fundamental units underlying AI compute. Forge virtually logs into clusters distributed worldwide and makes technical adjustments so that mismatched configurations work in concert, enabling traders to both transact in a standardized unit of compute and retain control over the underlying hardware.

Zhang, 31, describes the capability as functionally equivalent to physical presence inside a data center. He sees it as the foundational layer for "physically" settled GPU futures — contracts that, upon settlement, would hand owners direct control over a cluster with known performance characteristics.

That distinction matters because the alternative, cash-settled futures, carries significant basis risk. GPU architecture varies substantially depending on assembly configuration, and networking, memory, and security conditions all affect real-world output. At the same time, financing agreements frequently prohibit builders from disclosing prices, meaning a survey-based index may capture average setup costs while omitting outlier expenses — a gap that makes cash settlement an imperfect hedge for operators holding physical hardware.

"Our guess is when the cash-settled futures come out they're not going to be reflective of what's going on in reality," said Hong Chan Kim, Hyperbolic's head of finance and a veteran of Point72, Balyasny Asset Management, and Citadel LLC, the hedge fund. "Physical usage is what will tether that connection between the reality and the derivatives."

Hyperbolic has raised $20 million since its 2022 launch, is currently in the midst of an additional funding round, and expects to reach a $150 million annualized run rate by year-end. The company says it has approximately 250,000 users and roughly $200 million worth of annual GPU compute connected to its platform.

The startup is not alone in identifying the gap. CME Group is pursuing a cash-settled futures product tied to GPU pricing maintained by Silicon Data, an index provider backed by DRW, the Chicago-based proprietary trading firm. That contract is under review by the Commodity Futures Trading Commission, which recently extended its review period to November 9 — a timeline that also appears to affect a competing contract proposed by Intercontinental Exchange, backed by Andreessen Horowitz's a16z crypto fund, and data provider Ornn.

Don Wilson, founder and CEO of DRW, said investor appetite for the category is clear. "There's a ton of demand and capital flowing into this ecosystem so there's a lot of risk," Wilson said. "Cash futures are a great first building block. There are plenty of electricity futures that are cash-settled, oil futures that are cash-settled, so I think you can start these things either way."

Carmen Li, CEO of Silicon Data and DRW-backed Compute Exchange — a GPU marketplace also laying groundwork for physically settled futures — acknowledged the limits of index-based products. "Indices are a normalized global number," Li said. "You get financially hedged, but if you're looking for procurement, that's not going to help you."

Zhang framed the stakes in terms of what the current downturn in AI cloud stocks reveals about structural fragility. "CoreWeave's debt situation would be less concerning if providers had more ways to monetize available GPU capacity during a slowdown, especially when long-term contracts carry thin margins," he said. "Cash-settled futures can help falling prices, but they do not by themselves put idle GPUs to work."

With the CFTC review of competing products approaching a November deadline, the race to establish the dominant instrument for AI hardware risk is entering a critical phase — and whether physically settled or cash-settled contracts win out could shape how the broader industry manages the next cycle of AI volatility.

Disclaimer