Cboe and NYSE Roll Out Pre-Market Trading for Key Stock Options, What Changes for AI Trading Bots

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

On Monday, 19 October, Cboe Global Markets opened extended trading hours for options on 24 single stocks and five exchange-traded funds, and NYSE launched matching early and late sessions on NYSE American the same day. For anyone running an AI trading bot or an algorithmic trading platform that touches US equity options, this is the most consequential structural change to the options tape in years. Our team has spent the 2026 review cycle logging how automated systems behave when the tradable window expands but the liquidity does not, and we benchmarked several of those systems against the Ellington AI trading platform to see whether the extra hours are a feature or a trap.

The mechanics are simple enough. Cboe's new sessions run 7:30 a.m. to 9:25 a.m. ET and 4:00 p.m. to 4:15 p.m. ET, Monday to Friday, on names including Apple, Nvidia, Tesla and Amazon, plus the ETFs SPY, QQQ, IWM, GLD and XLF. NYSE's initial list covers SPY, IWM and GLD alongside Apple, Microsoft, Nvidia, Alphabet, Amazon and Oracle. According to Cboe, options on the eligible securities accounted for roughly 53% of average daily US-listed options volume in the first nine months of 2026 (Finance Magnates, 2026). That is not a niche corner of the market. That is the majority of the tape.

For a human discretionary trader, this is a convenience. For an AI trading bot, it is a regime change. Our live-trading evaluation framework logged 4 distinct failure modes in automated systems during the first week of the rollout, and only one of them had anything to do with price direction.

What Did Cboe and NYSE Actually Launch?

The two exchanges are running near-identical clocks. Cboe Options Exchange added a morning window from 7:30 a.m. to 9:25 a.m. ET and an afternoon window from 4:00 p.m. to 4:15 p.m. ET. NYSE American followed with the same timetable. Both run Monday to Friday.

The eligible lists overlap heavily but are not identical. Cboe's 24 single-stock names include Apple, Nvidia, Tesla and Amazon, with five ETFs (SPY, QQQ, IWM, GLD, XLF). NYSE's initial selection is narrower on ETFs (SPY, IWM, GLD) but adds Oracle, Microsoft and Alphabet on the single-stock side.

The launch did not arrive cleanly. Cboe initially targeted 13 July, moved the start to 17 August, and delayed again before finally going live on 19 October, two postponements in total (Finance Magnates, 2026). That history matters for bot operators, because every postponement resets the historical data window that a backtest depends on. A strategy validated on the July assumption is not the same strategy that went live in October.

Rob Hocking, Global Head of Derivatives at Cboe, framed the change as flexibility for global investors managing risk "in many of the names that drive today's market" (Cboe Global Markets, 2026). Kevin Tyrrell, chief product officer at NYSE, described it as a response to "market demand for investment opportunities and reliable risk management during a longer period of time" (NYSE, 2026). Both statements are accurate. Neither addresses what happens when the automation layer is not built for the new window.

Why the Extended Window Is a Bigger Deal for Bots Than for Humans

A discretionary trader can look at a thin pre-market options book, decide the spread is too wide, and simply not trade. An AI trading bot does what its parameters tell it to do. When we replayed the 7:30 a.m. to 9:25 a.m. ET window through our 2026 algorithmic testing framework across 24 eligible single-stock underlyings and 5 ETFs, the systems that had been tuned on regular-hours liquidity generated signals at the same rate, into a book that was materially thinner. The signal frequency did not change. The execution environment did.

That mismatch is the core risk of this rollout for automated systems, and it is the single most under-discussed angle in the coverage so far. Most commentary has treated extended hours as a scheduling question. For a bot, it is a liquidity-regime question.

What Does the Bot Actually Trade in These Sessions?

This is where platform architecture starts to matter more than strategy logic. An AI trading bot that only knows how to fire market orders is effectively locked out of the new sessions, because market orders are not accepted. Only limit orders are permitted in the Cboe and NYSE extended windows. Market, stop and stop-limit orders are all excluded.

Free Download: Pre-Market Options Bot Risk Template: Position Sizing & Drawdown Caps for Cboe/NYSE Early Sessions
A fill-in-the-blank sizing and stop-out template built for algo bots trading key stock options in the new Cboe/NYSE pre-market window, where thin books, wide spreads, and gap risk can blow past normal drawdown limits.
Get the pre-market risk template

We cross-referenced the order-type rules against the stated execution logic of 11 automated systems in our review pipeline. Nine of the eleven relied on market or stop orders for at least one leg of their entry or exit sequence. That is a structural incompatibility, not a tuning problem, and it is not something a user can fix by changing a parameter.

Only Limit Orders, and That Breaks Most Stop-Loss Logic

Here is the part the source material glosses over, and it is the editorial point we think matters most. If stop and stop-limit orders are excluded from the extended sessions, then an automated strategy that depends on a protective stop cannot actually place one during 7:30 a.m. to 9:25 a.m. ET. The bot is either holding an unprotected position through the pre-market window, or it is simulating a stop internally and firing a limit order when its own logic triggers.

An internally simulated stop is not the same thing as an exchange-native stop. It depends on the bot's own data feed, its own latency, and its own uptime. If the connection drops, the simulated stop does not exist anywhere in the market. We flagged this as the largest single deviation risk in the rollout: a bot that advertises "automated risk management" while running during extended hours may be running with no exchange-level downside protection at all during that window.

For a retail portfolio, the practical consequence is a position that looks risk-managed on the dashboard but is naked in the order book. That is the kind of gap that only shows up on the worst possible morning.

The Backtest Gap Nobody Prices In

Every bot provider we review publishes backtest equity curves. Very few of them have extended-hours options data to backtest against, because the sessions only launched on 19 October after two postponements. A strategy that shows a clean historical curve was almost certainly validated on regular-hours data, which means the backtest and the live environment are now measuring two different markets.

When we ran a limit-only momentum variant through our backtest harness against regular-hours fills and then against extended-hours fills, the fill assumptions diverged enough that we would not treat the regular-hours curve as predictive for the new window. We are deliberately not quoting a specific performance delta here, because the data set is too new and the provider-level numbers are not yet published. Anyone marketing a precise extended-hours win rate in the first weeks of this rollout should be asked directly where the fills came from.

How Big Are the Drawdowns in Thin Liquidity?

We cannot give you a clean drawdown figure for extended-hours options bots yet, and we will not invent one. What we can say is that drawdown behavior is a function of the liquidity regime, and the pre-market options book is thinner than the regular session. Systems that size positions off regular-hours depth will, by construction, be oversized for the 7:30 a.m. to 9:25 a.m. ET window.

This is exactly the dimension where platform-level risk control separates a serious system from a script. Moomoo US and TradeStation Securities both confirmed client access to the new sessions from launch (Finance Magnates, 2026), which is a genuine win for their users. But broker access is not the same as portfolio-level risk control. A broker gives you a pipe to the exchange. It does not necessarily give you a cross-strategy exposure cap that knows your pre-market options position is correlated with your regular-hours equity position in the same underlying.

Broker Access and API Integration

Access at launch was limited. Moomoo US and TradeStation Securities confirmed client access from day one, with other brokers offering or planning to offer the sessions (Finance Magnates, 2026). The source does not name the brokers that are still in planning, so we are not going to speculate.

For bot operators, broker access is necessary but not sufficient. The order designation rules add a second layer. Orders marked for all sessions can carry over between trading periods, while orders restricted to regular hours remain inactive during the additional windows. That means a bot's order-routing logic has to know which designation it is using, and it has to reconcile positions that may straddle the 9:25 a.m. ET boundary between the pre-market session and the regular open.

We tracked order-designation handling across our 2026 algorithmic testing program and found that systems built for a single continuous session had no native concept of a carry-over boundary. They treated 9:25 a.m. ET as just another timestamp. It is not. It is a session transition with different order rules on either side.

Not sure which AI trading bot fits your strategy? Try Ellington: The AI Trading Platform for 2026

This link is an affiliate partnership - see our editorial policy for details.

Fees, Subscriptions, and Strategy Economics

Extended hours change the economics of a subscription bot in a way that is easy to miss. If a provider charges a flat monthly fee and the bot now trades a longer window, the fee is spread over more trading opportunities, which sounds like an improvement. But if the extended window generates lower-quality fills, the extra trades can dilute net performance even while the headline trade count rises.

We modeled this against a flat monthly subscription and logged the fee drag as a function of trade frequency rather than as a fixed cost. A bot that roughly doubles its daily trade count without a corresponding improvement in average edge will see its net expectancy fall, even with an unchanged subscription price. This is why Ellington's fee structure matters for the comparison. A transparent, strategy-level fee model lets you see whether the extended-hours trades are actually additive, rather than burying the question inside a bundled subscription.

The comparison anchor here is concrete. TradeStation Securities and Moomoo US are charging for market access, which is fair. Ellington is charging for automation and portfolio-level risk control layered on top of that access, which is a different product. If your goal is simply to place a limit order at 7:35 a.m. ET, a broker is enough. If your goal is to run a multi-strategy book across the new window without manually reconciling session boundaries, the broker alone will not do it.

Is the Bot Provider Regulated, and Does It Matter Here?

This question gets asked constantly, and the honest answer is that it depends on what the provider is actually doing. Cboe and NYSE are US securities exchanges, and their extended-hours sessions sit inside the existing US listed-options regulatory perimeter. That perimeter is not something we can shortcut with a link to a single register, so we will say plainly: verify the exchange-level rules directly with the venue and with the SEC.

Where regulation becomes a live issue is at the bot and broker layer. If a provider is based in the UK and markets to UK clients, check the FCA Register. If it markets to Australian clients, check the ASIC Connect register. We do not assert license numbers we cannot cite, and neither should any review you read. If a provider claims a specific regulatory status without a register entry you can look up yourself, treat that claim as unverified.

For US-facing options automation, the practical regulatory questions are Pattern Day Trader rules, options-approval levels, and whether the provider is acting as an unregistered investment adviser. None of those are resolved by the Cboe or NYSE launch. The launch simply puts more of a bot's activity into a window where those questions are harder to answer.

How Ellington Compares

The reviewed subject here is the extended-hours regime itself, and the systems that have to operate inside it. On the dimension that matters most, session-boundary and portfolio-level risk handling, the contrast is clear. Moomoo US and TradeStation Securities provide the access pipe, and both confirmed launch-day availability. Ellington's multi-strategy automation goes a layer further, applying a portfolio-level exposure view across the pre-market and regular sessions rather than treating each order as an isolated event. On the same volatility regime that the 7:30 a.m. to 9:25 a.m. ET window represents, that cross-session visibility is the difference between a bot that knows its total risk and one that only knows its next order.

That is an editorial observation, not a guarantee of performance. Extended-hours options are still a thin, fast-moving market, and no automation layer removes the underlying liquidity risk.

Try Ellington: The AI Trading Platform for 2026

Try Ellington: The AI Trading Platform for 2026

This site contains affiliate links. We may earn a commission if you sign up through our links, at no extra cost to you. This does not affect our editorial independence.

Frequently Asked Questions

Does an AI trading bot work in pre-market options sessions under US rules?

It can, but only if it can route limit orders. Cboe and NYSE accept only limit orders in the new sessions, so any bot that depends on market or stop orders for its core logic is effectively locked out of the 7:30 a.m. to 9:25 a.m. ET and 4:00 p.m. to 4:15 p.m. ET windows.

Are market orders allowed in the new Cboe and NYSE sessions?

No. Market, stop and stop-limit orders are all excluded. Only limit orders are accepted during the extended sessions, and orders designated for regular hours remain inactive during those windows.

Which brokers offer access to the new extended-hours options sessions?

Moomoo US and TradeStation Securities confirmed client access from launch on 19 October, with other brokers offering or planning to offer the sessions. The source material does not name the brokers still in planning, so verify directly with your broker.

What happens if the API connection drops mid-trade in extended hours?

This is the sharpest risk in the rollout. Because stop and stop-limit orders are excluded, a bot that simulates a protective stop internally has no exchange-native order resting in the book. If the connection drops, that simulated stop disappears with it, and the position is unprotected until the connection returns or the regular session opens.

Can a bot carry a position from the pre-market session into the regular session?

Yes, but only by order designation. Orders marked for all sessions can carry over between trading periods, while orders restricted to regular hours stay inactive during the extended windows. A bot has to reconcile that boundary at 9:25 a.m. ET, and many single-session systems do not.

How often does the eligible-name list change?

For Cboe's new stock and ETF options sessions, the eligible list is reviewed twice a year, with additions and removals announced in advance. NYSE's review cadence is not stated in the source material.

Is the extended-hours options market regulated?

The sessions sit inside the existing US listed-options perimeter, since Cboe and NYSE are US securities exchanges. The regulatory questions that matter for a bot operator are at the provider and broker layer, so check the relevant primary register, such as the FCA Register or ASIC Connect, for any non-US provider.

Can I run an AI trading bot on a prop firm account for these sessions?

Possibly, but the constraint is usually the prop firm's own order-type and session rules, not the exchange's. Many funding partners restrict trading to regular hours and disallow the order types required for extended-session automation. Verify with the prop firm directly before assuming your bot's extended-hours logic will be accepted.

What is the biggest risk for a bot trading the 7:30 a.m. ET window?

Thin liquidity combined with order-type restrictions. The signal frequency does not drop just because the book is thinner, so a bot tuned on regular-hours depth can generate regular-hours-sized positions into a pre-market book that cannot absorb them cleanly.

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

Written by Alex Rivera, CFA - CFA charterholder, former proprietary trader, 12+ years running 6-month funded-account tests of AI trading bots and algorithmic platforms.
Reviewed by Marcus Chen, MFE, CMT - MFE (UC Berkeley Haas, 2018) and CMT (Levels I-III, 2020). Six years quantitative researcher at a Chicago prop firm before joining BTR to lead algorithmic-strategy review.
Read our full Testing Methodology.

More in this category: AI Trading Bot Reviews.