85% of Crypto Traders Use AI Trading Tools and 6.5% Would…

Almost every crypto trader in MEXC’s new AI trading survey has used the technology, and almost none of them will let it near the money. The exchange’s AI Trading Report, published on Wednesday, found 84.8% of 3,708 respondents had used AI to research or trade crypto in the previous six months, while 6.5% would authorize it to trade more than half their account balance.

One methodological caveat governs how every number in the report below should be read. Respondents were recruited through MEXC’s own social media, app and website between 14 and 20 September, and skew male and toward active spot and futures traders, so this is a self-selected sample of one exchange’s users rather than a cross-section of the market.

84.8% Use AI, 6.5% Would Hand It More Than Half the Account

The AI trading numbers fall away at every step along the path toward autonomy. Usage sits at 84.8%, and 65.9% act on AI recommendations at least sometimes by placing orders or adjusting positions. Only 20% have tried execution without confirming each trade, and 13% have connected an AI agent to an exchange account at all.

Each step down the chart hands the AI more control, and the share of traders willing to take it drops by an order of magnitude. Source: MEXC AI Trading Report, October 2026 · Chart: FinanceFeeds

On how much capital they would commit, with transfers and withdrawals disabled and the ability to stop trading at any time, 54% would allow AI to use some funds but no more than 20% of the account. Another 18.3% would allow none. The 6.5% who would go past half are the outliers, and 57.6% said their own judgment with AI assistance worked better than either extreme.

58.3% Report Better Research, While 23.1% Say Execution Got Worse

The reported AI trading benefit is concentrated before the order rather than at it. Among AI users, 68.9% said the technology had its biggest impact in research, opportunity discovery, strategy and risk management, and 58.3% said it improved their market research.

Execution is where the reported picture turns against the technology, and sharply so. Only 46% reported better trade execution, 30.9% saw no difference, and 23.1% said their results had worsened. That split sits awkwardly beside the enthusiasm, and it helps explain why so few traders will delegate the order itself.

Tool choice in AI trading shows a gap the exchanges will read closely. General-purpose assistants had been used by 76.6% of AI users, against 38.7% for exchange chat assistants and 17.1% for external agents connected to accounts, yet 62.6% said they would prefer AI built into an exchange. That is also the finding most useful to the company that commissioned the survey, which is worth noting even though the direction is plausible.

Investor Takeaway

Caps on Funds and Losses Move AI Trading Willingness From 55% to 71.8%

Hesitation about AI trading softened sharply once respondents could set the boundaries in advance. Under a baseline scenario offering notifications and the option to pause or revoke access, 55% were willing to let AI execute automatically. Adding limits on funds, positions, leverage and losses, with trading halted automatically at the loss threshold, lifted that to 71.8%.

Hard limits move the average trader 16.8 points, and move the already-experienced close to four in five. Source: MEXC AI Trading Report, October 2026 · Chart: FinanceFeeds

More than half of those initially unwilling or uncertain, 53.6%, changed their answer once the limits were described. Among traders who had already let AI place orders subject to their approval, willingness reached 77.8%, near the 78.9% recorded by those who had tried automatic execution. The movement runs both ways, since 13.3% of those willing under the baseline became hesitant once limits entered the picture.

Vugar Usi Zade, chief executive of MEXC, read that as a demand for boundaries rather than for less automation in the comment issued with the report.

"As AI becomes more integrated into trading, our findings show that users don’t want less control. They want clearer boundaries. For traders, that means clear risk limits, the ability to intervene when needed, and confidence that the platform has done its part before automation begins."

Hacken’s Attack Scenario Explains Why 95.5% Want a Mandatory Safeguard

Hacken, the blockchain security firm, contributed the technical analysis, which is separate from the survey findings and is not an audit of MEXC’s infrastructure. Its central point is that the protection AI trading users reach for first does not cover the risk they actually face. Disabling withdrawals stops an agent moving funds out, but an attacker can still manipulate it into buying a low-volume token at an inflated price from the attacker’s own sell orders, using the trading permissions the agent already holds.

Hacken also described a lab test in which an instruction embedded in a software changelog caused an agent to run a script without explicit approval. No trading was involved, though the point carries over to news pages and token descriptions an agent might read while researching. Its recommendations are to check order prices against an independent reference, and to set trading limits in a separate system the agent cannot override, because several orders each under a $10,000 cap can still build a position larger than the user intended.

Traders who have actually let AI execute are the ones asking hardest for audits and incident disclosure. Source: MEXC AI Trading Report, October 2026 · Chart: FinanceFeeds

Traders broadly want to set their own terms, with 46.2% preferring that exchanges supply safety settings and let users switch them on. The qualification is large, because among that group 97.8% still backed at least one mandatory safeguard, and across all 3,464 valid responses to the standards question support reached 95.5%. Activity logs ranked last among user priorities even though 51% of automatic-execution users wanted incident investigation, which is the same record viewed from the other end.

MEXC has run this kind of research before, including an August report with CoinGecko that found 74.2% of traditional finance users had shifted trading to crypto exchanges, and it has partnered on AI-driven identity fraud detection with Sumsub. The contrast with institutions is the useful one, because European banks and asset managers have moved AI into full-scale production in regulated workflows while retail traders hold the order button.

Investor Takeaway