October 21, 2026 Trader working at the New York Stock Exchange.
new york stock exchange
as coreweave With stock prices plummeting and AI complexes weighing heavily, investors are faced with a simple question. What is the best way to hedge AI risks?
A former Citadel quant may have the answer.
Jasper Zhang is a former quantitative researcher at Citadel Securities and currently leads Hyperbolic. The company is a 50-person financial technology startup building a spot market for tradable GPU computing power, a tool that allows traders to avoid falling AI demand.
The platform allows investors to buy and sell GPU clusters, the units that power AI computing. This is done through Hyperbolic’s Forge software. The software virtually logs into clusters around the world, makes technical adjustments to help mismatched clusters work together better, and allows traders to not only buy and sell standardized units of compute, but also control the underlying GPUs for their own purposes.
Zhang said this is as good as it gets in the data center itself, and believes the technology is the first building block for “physically” settled GPU futures, and once settled, the contract holder will have full control over a cluster with known characteristics. In other words, investors have a new way to hedge against declining demand for AI by trading the hardware itself.
Today, GPU architectures vary widely depending on how they are assembled, resulting in varying levels of output. It varies by data center. Networking, memory, and security all affect GPU performance. At the same time, financing agreements often prevent builders from sharing prices. This means that survey-based metrics average out setup costs and may not include outlier costs.
In the world of futures, this is called “basis risk.” This is a problem for market makers to solve, and Zhang hopes Hyperbolic can fill the void.
Mr. Zhang, 31, came to the United States with some Chinese math medals, completed a PhD in mathematics from the University of California, Berkeley in just two years, and took a job at Citadel Securities, which builds trading models. Hyperbolic’s head of finance, Hon Chan Kim, is a hedge fund veteran who worked at Point72, Balyasny Asset Management, and hedge fund Citadel LLC.
Hyperbolic has raised $20 million since its founding in 2022 and is in the midst of another round of funding, with an expected run rate of $150 million annually by the end of the year. The company says it currently has about 250,000 users and an estimated $200 million worth of GPU computing per year connected to the platform.
emerging market
Other major companies see similar opportunities in the market for AI hedging tools.
CME Group plans to offer a cash-settled futures product tied to GPU pricing maintained by Silicon Data, an index provider backed by DRW, one of Chicago’s largest prop trading firms.
“There’s a lot of demand and capital coming into this ecosystem, so there’s a lot of risk,” DRW founder and CEO Don Wilson said by phone. “Cash futures are a great first building block. There are a lot of cash-settled electricity futures and cash-settled oil futures, so I think you can start with these either way.”
CME products are currently under review by the CFTC, which recently extended that review period until Nov. 9, and will likely conclude another futures contract proposed by CME. intercontinental exchange and Orn, a data provider backed by Andreessen Horowitz’s a16z crypto fund.
“Our guess is that the introduction of cash-settled futures will not reflect what is actually happening,” Hyperbolic’s Kim said. “Physical usage is what connects the connection between reality and derivatives.”
The bigger problem is that physical futures are imperfect hedges against real-world risks. If AI demand hits a wall, AI rental centers like CoreWeave could be left with billions of dollars worth of AI hardware with little way to turn it into cash.
“The index is a normalized global number,” said Carmen Li, CEO of Computing Exchange, a GPU buying and selling marketplace backed by Silicon Data and DRW. The computing exchange is also laying the foundation for physically settled futures. “You can hedge financially, but if you’re looking for procurement, that doesn’t help.”
Zhang argues that what is most needed in the AI complex, especially small startups that don’t have the capital or credit to take the risk of purchasing large amounts of hardware, is a virtual pipeline to physical hardware.
“CoreWeave’s debt situation would be less of a concern if providers had more means to monetize available GPU capacity during economic downturns, especially when margins are thin on long-term contracts,” Zhang said. “Cash-settled futures can help drive prices down, but by themselves they can’t bring idle GPUs to life.”
