NYSE owner plans compute futures as AI capacity becomes a tradable risk

Intercontinental Exchange, the owner of the New York Stock Exchange, wants to create futures contracts for computing power. That is a small move inside market infrastructure, but a big signal for every company buying, building or committing to AI capacity.

Bloomberg reported Tuesday that ICE is working with financial-infrastructure firm Ornn. The plan is to use Ornn indexes that track GPU computing costs as the basis for new futures contracts. The contracts would be denominated in US dollars, settle in cash and reference indexes covering several major GPU types. The plans are still subject to regulatory approval.

This may sound narrow. It is not. AI compute is starting to acquire the mechanics of a commodity market: reference pricing, risk transfer and tools for hedging price swings. ICE is not alone. Bloomberg also noted that CME Group announced its own plans for compute futures the week before. When two major venues move in the same direction, the story is not a quirky new derivative. The story is that compute cost has become large, volatile and strategic enough to need market infrastructure.

For executives, the practical consequence is simple. AI budgets are no longer only about software licences and cloud consumption. They are about capacity, contract duration, price risk and supplier power. A company signing large commitments for GPU capacity, model use or AI platforms is effectively taking exposure to a market shaped by chip supply, energy costs, data-center buildout, cloud pricing and frontier-model demand.

That matters most for organizations putting AI into core work: banking, manufacturing, energy, telecoms, public services and software development. As agents, inference volume and data demands grow, compute can become an operational input like electricity, network capacity or capital. CIOs cannot manage that alone. CFOs, procurement, risk teams and boards need to understand how the cost can move.

Bloomberg quoted Trabue Bland, ICE’s senior vice president of futures markets, saying that compute has evolved quickly and needs a globally accepted pricing mechanism and risk-management tool. Ornn CEO Kush Bavaria said compute has grown into a trillion-dollar market but still lacks the pricing and risk-transfer infrastructure relied on by other major commodities. That is vendor language, but the point is real. AI capacity is often priced through cloud contracts, reserved instances, private discounts, model APIs and long infrastructure commitments. That makes comparison and governance hard.

A more visible compute price could change enterprise negotiations. Large buyers may start demanding price protection, index-linked terms, exit clauses and clearer disclosure of how AI vendors pass through capacity costs. Data-center operators may use the instruments to manage revenue exposure. Cloud customers may finally get a cleaner way to ask whether a three-year AI agreement reduces risk or simply buries it in the contract.

This does not mean most companies should trade compute futures. They should not turn AI procurement into speculation. But large AI buyers should track the market. If compute gets a standardized price, it becomes easier to see whether a contract is expensive, cheap or risky. It also becomes easier to test whether local deployment, model switching, caching strategies, smaller models or diversified suppliers actually improve resilience.

The board question is blunt: what happens to the AI plan if compute prices double, if a chosen supplier tightens capacity, or if regulation forces more processing into specific regions? The answer should not be that the company hopes the cloud bill behaves.

The deeper point is that AI is leaving the innovation budget. When the main input behind AI gets a futures market, it moves into finance, procurement and operational risk management. That is where leadership teams should place the discussion now.

Sources and media

• Primary source: Bloomberg, "NYSE’s Owner Plans Its Own Futures Market for Computing Power", published May 19, 2026: https://www.bloomberg.com/news/articles/2026-05-19/nyse-s-owner-plans-its-own-futures-market-for-computing-power
• Company context: Ornn describes its OCPI GPU-compute index and cash-settled compute contracts at ornn.com.
• Bloomberg’s article image was used only as source/right context, not rehosted.
• Thumbnail: OpenAI Image 2 / hogby.ai.