Why is AI trading accelerating its focus on the futures market?

On March 3, Michael Selig, chairman of the Commodity Futures Trading Commission (CFTC), stated at the Milken Institute’s “Future of Finance” conference that the CFTC will introduce a regulatory framework for cryptocurrency perpetual contracts within weeks, aiming to gradually bring this trading product, which has been largely dominated by offshore exchanges, back to the U.S. domestic market. This statement is a continuation of the U.S. market’s ongoing efforts in this area over the past year. In July 2025, Coinbase launched CFTC-regulated perpetual futures products for U.S. retail users; in December 2025, Cboe listed continuous futures products for Bitcoin and Ethereum; by March 2026, Coinbase further expanded its product line for non-U.S. users with stock perpetual futures. It can be seen that perpetual futures are gradually becoming the core infrastructure for derivatives trading execution, and the U.S. is accelerating to catch up in this area.

AI trading is often marketed as a smarter way to trade cryptocurrencies. However, when focusing on practical applications, it is actually better suited for the futures market. Futures contracts inherently possess characteristics such as standardization, margin-driven mechanics, daily mark-to-market, and a more symmetrical structure for both long and short positions, making systematic execution easier to implement compared to the spot market. The logic of spot trading often gets entangled with a series of non-trading operational issues such as custody, settlement, and borrowing mechanisms, which can vary greatly between platforms (if you want to short). Futures eliminate these burdens. The capital and strategies for automated trading are increasingly concentrated in the derivatives market, where perpetual contracts account for the vast majority of trading volume in crypto derivatives, making this trend unsurprising.

Retail investors are rapidly moving from copying trades and signals to automated execution. People who previously copied calls in Telegram groups are now starting to subscribe to trading bots, and some even begin to build their own systematic strategies. The built-in margin mechanism and contract-level standardization of the futures market make this transition the easiest to realize.

What the futures market offers to machines, the spot market cannot provide.

Spot trading means holding assets directly. Even in an exchange with clear matching rules, price priority, and time priority, the algorithms have to deal with custody, settlement, and vastly different borrowing mechanisms due to platform differences (if you want to short).

Futures contracts extract these steps from the trading logic. Based on margins, daily mark-to-market, and natural symmetry between longs and shorts, the same strategy can express views in both directions. Position size becomes an adjustable parameter linked to margin, and risk limits directly correspond to margin thresholds. The model’s adjustments in risk control and position management are finer-grained, and the parameters are clearer.

For automated strategies, this difference directly alters the methods of risk management, position calculation, and execution. The regulatory framework considers margin and daily mark-to-market as fundamental mechanisms of the futures market, manifested in standardized terms, centralized clearing, margin as performance guarantees, and daily settlements. These mechanisms provide liquidity and scalability for the futures market, while also making it easier to be transformed into a rules-based trading system.

Perpetual contracts do not have an expiration date. The funding rate (usually settled every eight hours) serves an anchoring function, pulling the price of perpetual contracts back towards the spot price. The calculation of the rate is based on the recent price difference between futures and spot. For systematic strategies, the funding rate is an additional state variable. It reflects in real-time the position bias and leverage distribution between longs and shorts. Such signals are not available in the spot market.

Signals that only exist in the derivatives market.

The data layer generated by the futures market does not exist in the spot order book. This is the most underestimated reason why automated trading leans towards derivatives.

Basis (the price difference between spot and futures) and funding rate (the cash flows periodically paid between longs and shorts in perpetual contracts) are important signals for judging the degree of deviation and leverage direction in the derivatives market. They inform models how far derivatives deviate from the underlying asset and which direction the leverage is leaning. Models can treat this deviation as feature input, risk control signals, or a combination of both.

Open interest provides a second layer of market intent information. When perpetual contracts dominate both the transaction volume and open interest in Bitcoin futures, the embedded position information in derivatives is the densest in the entire market. Micro-structural patterns, clearing cascades, and sentiment proxy indicators often first emerge in the futures market, as participants express their judgments through leveraged funds in futures. For models, the place with the densest signals is often the most worthwhile to learn.

The same applies to execution. The standardized contract specifications in the futures order book, clear matching rules, and granular order book data are inherently suitable for machine learning. Execution optimization and order book modeling are machine learning applications that coexist with market structure in the derivatives market. In the spot framework, they resemble a later added auxiliary capability.

Why price discovery matters for automated trading.

Another often underestimated advantage is that futures typically dominate price discovery.

Studies on the dynamics of spot and futures prices repeatedly show that under normal market conditions, futures contribute the majority of price discovery. When arbitrage signals appear, this proportion further expands. In the cryptocurrency market, standard price discovery indicators point to futures dominance. The deviation between futures and spot can predict subsequent movements in spot, but the reverse does not hold. Information tends to first reflect in futures, then transmit to spot, with a time lag in between.

The foreign exchange market provides a useful reference. During periods of lower transparency in the spot market, futures exhibit a disproportionate amount of information content, sometimes leading the spot market by minutes. After transparency in the spot market improves, information flow gradually returns to spot, with market design and transparency determining where informed capital concentrates. Futures trading venues, as centralized and rules-driven bidding environments, have machine-readable transparency, naturally attracting this kind of capital. For systematic models, learning the mapping relationship between market conditions and trading actions is cleaner where signals are concentrated.

Being better for AI does not mean being safer for everyone.

Futures compress time. Leverage amplifies both gains and losses. Margin serves as a performance guarantee; when an account falls below the maintenance margin level, traders must add variation margin. In crypto perpetual contracts, the contract itself is a high-leverage tool, and the details of order protection (e.g., when the latest contract price deviates from a reasonable benchmark price beyond a threshold, orders triggered by stop-loss or take-profit will be rejected) directly affect the execution results of any robots operating in that venue.

Several factors are non-negotiable for automated systems. Assumptions about slippage must be conservative, operational monitoring must be continuous, and the perception of margin models must be clear. A position may be liquidated even when there is capital elsewhere on the platform, depending on whether isolated margin or cross margin is used at that time. These risks do not disappear because the executor is an algorithm. Systems designed around them can contain risks. Systems that ignore them will ultimately be bitten back by amplified risks.

What AI really needs is structure; predictive capability is only part of it. Structure means knowing how it will operate even when the market is disordered.

What this means.

The structural fit between automated strategies and the futures market is giving rise to a new class of native futures trading platforms. These platforms are built around derivatives infrastructure from the outset, with automated capabilities embedded within the trading architecture.

OneBullEx is an example of this approach. Its 300 SPARTANS operate directly on proprietary futures infrastructure, with net value and historical performance being traceable and auditable. OneALPHA transforms natural language inputs into deployable futures strategies, allowing non-coding users to enter systematic trading. If the market itself has already provided the standardization, signals, and risk architecture needed for systematic strategies, then the platform should be built around such structure from day one.

More important than any single platform is the overall trend. AI-native trading is most likely to mature first in the futures market because futures are inherently built for structured execution.

AI will continue to evolve, but the discipline it truly needs is not a new invention. The futures market is designed for such discipline.