Around May 20, 2026, two seemingly unrelated pieces of news emerged almost simultaneously from the financial world: on one side, the parent company of the New York Stock Exchange, the well-established derivatives operator Intercontinental Exchange Group (ICE), announced plans to collaborate with financial infrastructure company Ornn to write "costs of computing power such as GPU" into futures contract terms, to be traded on regulated exchanges, with the product itself awaiting regulatory approval; on the other side, AI fintech company Moment, founded by former quantitative trading and research personnel from Citadel Securities, announced the completion of a new funding round of $78 million, led by Index Ventures, with existing shareholders a16z and Avra continuing to increase their stakes, subsequently connecting quickly with traditional wealth management and brokerage giants like Edward Jones and LPL Financial to integrate their AI platform into the daily decision-making processes faced by end clients. In this same time window, one is abstracting the cost of computing power supporting large models into a futures asset that can be hedged or speculated, while the other is turning generative AI into a "standard tool" within investment research and wealth management scenarios, linking traders, advisors, and individual accounts. The common direction behind this set of coincidences is very clear: in the eyes of Wall Street, AI is no longer merely "the story of which company to invest in," but is being broken down into measurable computing resources and packageable service capabilities, gradually being incorporated into the toolbox of financial engineering, and beginning to move towards being priced, leveraged, and financialized.

GPU computing power included in futures contracts

ICE and Ornn aim to abstract the buzzing GPUs in the data centers into a price curve that can be monitored. According to current disclosed information, ICE plans to list a futures contract based on a "trading-type computing power index" provided by Ornn on its regulated exchange, with the objective of this index being straightforward: to track changes in the cost of computing hardware, such as GPUs, that support the current AI industry. Ornn positions itself as a "financial infrastructure company," responsible for compressing the computing power prices scattered in the equipment procurement, custody, and leasing markets into a number that can be recognized by trading systems, while ICE incorporates this number into a standardized contract and brings it to the main board that has been operated for many years. For market participants, this means that GPU costs are appearing for the first time in a compliant, hedgeable form alongside quotes for crude oil and interest rates; however, this contract is currently still awaiting approval from regulatory agencies, and specifics regarding the launch timetable, settlement method, contract size, and trading rules have not been disclosed.

Once these types of computing power futures are implemented, the protagonists of the story will no longer just be chip manufacturers but all buyers and sellers who are "choked" by GPU costs. AI companies training large models and cloud service providers purchasing graphics cards in bulk can lock in a certain index level for a future period on the futures market to hedge against the risk of suddenly rising GPU costs, giving their financial statements a reference for hardware expenses that can now be hedged; on the other side, those providing computing power custody or resale services can utilize the same contract to hedge against the uncertainty of reduced income caused by future price declines. More importantly, once a regulated main board standardizes the issue of "computing power costs," the market will be forced to draw a forward curve across time periods, with valuations for AI projects, long-term procurement contracts, and even merger negotiations all gaining a "public price" to refer to. GPU will no longer be just a cost item and technical detail, but will start to become a financial variable that can be systematically priced and managed by Wall Street.

From oil to computing: Wall Street bets on a new resource

In the narrative of Wall Street, "AI computing power is the new oil" is transforming from a slogan into a tradable asset. Over the past few decades, oil, natural gas, and electricity have been financialized not because they are sufficiently "cool," but because the underlying resources have been standardized into contracts: oil priced by the barrel, electricity priced by the kilowatt-hour, and then listed on regulated exchanges to trade forward prices and manage price risks. ICE and Ornn are replicating the same path—first abstracting the costs of computing power such as GPUs that support large model training and inference into a trading-type computing power index and then designing futures contracts on the main board, moving what was originally written on cloud service invoices into a financial coordinate system where it can be longed or shorted.

The reason computing power has the ground for financialization is that it possesses two key characteristics of "commodities": first, there is a highly homogeneous demand, and the recent explosion of large models and generative AI has caused almost all AI companies to scramble for the same batch of GPUs; second, there is significant price volatility, where tense supply and demand can rapidly erode the gross margins of a whole group of startup companies. For those AI companies where computing power is a core cost item, any uncontrollable fluctuation in computing power prices will directly affect cash flow and valuation models, while chip manufacturers and cloud providers are similarly exposed to pressures from upstream supply and downstream demand. Who will step into the "trading tunnel" of computing power futures? The most direct participants are the cloud service providers and leading AI model companies that are purchasing GPUs in large quantities, as they have the motivation to lock in training and inference costs for the next several years using futures; some participants in the upstream chip ecosystem can use futures to hedge against the risk of price declines, managing their production expansion bets and market volatility separately; while the entry of pure financial capital may not only serve as a hedge but also add a layer of liquidity and speculative premium to the computing power curve. As this mechanism takes shape, computing power costs will no longer just be a "passively accepted" expense but will gradually evolve into a bargaining chip for pricing power contested by participants in the AI value chain. Those who can influence and leverage the forward curve of computing power described by institutions like ICE are more likely to take the initiative in the next round of AI investment financing and merger negotiations.

Citadel's quant arm shifts to AI wealth management

On the other end of mapping computing power into forward curves lies the rewriting of wealth management interfaces themselves. The story of Moment began with a group of people engaged in high-frequency and quantitative research "leaving" Citadel Securities: they packaged years’ worth of experience in monitoring and matching trades, as well as building models to profit from millisecond advantages, into an AI fintech company that no longer does price spreads for proprietary funds but instead directs technology towards investment and wealth management scenarios, attempting to use models to predict and optimize traditional "human brain business" such as investment advising and asset allocation.

Capital quickly provided a valuation. Recently, Moment completed a new funding round of $78 million, led by Index Ventures, with original shareholders Andreessen Horowitz (a16z) and Avra continuing to make bets, with the list of investors almost being a "Silicon Valley + Wall Street" standard combination. The money behind this not only provides financial ammunition but also enhances channels and discourse power—Moment has already secured collaborations with large wealth management or brokerage institutions like Edward Jones and LPL Financial, embedding its AI capabilities into the investment advisory tools and asset allocation processes of the counterparties, as traditional institutions start using models to assist in assessing client risk preferences, product matching, and portfolio adjustments. In contrast to ICE's attempt to create GPU costs into futures contracts, on one end, computing power itself is being financialized into a tradable and hedgeable cost curve, while on the other end, the quantitative arm from Citadel is embedding AI directly into front-end wealth management services. The convergence of these two lines constitutes Wall Street's dual strategy of competing for both "computing power pricing power" and "dominance of wealth management interfaces."

Regulatory red lines and explorations of computing power futures

To bring GPU cost curves to regulated exchanges, ICE must first wait for regulatory approval. Currently, its collaboration with Ornn on computing power futures remains at the level of "plans" and "intentions," with public information lacking a clear launch timetable, not disclosing whether it will be cash-settled or physically delivered, and no visibility of contract size, margin ratios, or price fluctuation limits among other trading details. Under the regulatory framework for futures and derivatives in the U.S., any new form of commodity or index futures must first undergo a round of validation regarding "whether it is suitable to serve as a regulated asset," which means that even though exchanges and market institutions are eager, the product itself has a lot of uncertainties regarding its rollout.

For regulatory authorities, the focus on computing power futures is not just about whether they are "new," but "can they be manipulated." For the trading-type computing power index provided by Ornn to become a futures asset, it must clearly explain: where the constituent data of the index comes from, whether it is sufficiently diverse and transparent, whether there are any structural weaknesses that allow a single supplier or trading platform to manipulate quotes, and whether there are any conflicts of interest between the index developers and potential trading participants. Historical experience shows that regulators for such new index futures often require proof of a genuine economic hedging need, while also being challenging to be "hunted" by a few institutions. ICE has long dealt with regulators in the fields of commodities and financial derivatives, is familiar with compliance design pathways, and understands that it is presently probing an undefined red line: on one hand, there is a desire to be the first to occupy the new asset class of "computing power," setting industry standards ahead of potential competitors, while on the other, regulators must find a balance between encouraging financial innovation and preventing systemic risks. Whether and in what form computing power futures can be approved will itself become the first litmus test of the success or failure of this balancing act.

The next battleground after the financialization of computing power

From ICE's GPU computing power futures to Moment securing $78 million in financing and integrating into traditional wealth management channels, these two pieces of news complete the narrative: Wall Street is no longer satisfied with betting on "the next leading AI company," but is beginning to directly dismantle the computing power and AI capabilities that support these companies, turning them into financial components that can be traded and packaged into wealth management products. The next battleground will likely revolve around who gets to define the standards of these components—more derivatives anchored by costs of computing power, model usage rights, and service rates, as well as "financial + AI" services centered around AI investment advising and risk management tools will open a new round of competition among exchanges and various institutions: who controls the indices and data, who first educates a sufficient number of end users, and who can scale within compliant frameworks. For ordinary participants, what needs to be closely monitored are not promotional phrases, but a few cold hard variables: how computing power futures will ultimately be approved, the composition and long-term performance evolution of real users of AI wealth management tools, and whether the early products’ liquidity is sufficient to support complex strategies, because in a new track that is still in a period of exploration, pricing errors and compliance gaps often arise more quickly and fatally than the technologies themselves.

Join our community to discuss and improve together!
Official Telegram community: https://t.me/aicoincn
AiCoin Chinese Twitter: https://x.com/AiCoinzh
OKX benefits group: https://aicoin.com/link/chat?cid=l61eM4owQ
Binance benefits group: https://aicoin.com/link/chat?cid=ynr7d1P6Z

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。