AI bots capture easy money from prediction market glitches

A fully automated trading bot recently captured nearly $150,000 in profit by exploiting pricing glitches on short-term crypto prediction markets. The strategy relied on 8,894 executions that targeted five-minute Bitcoin and Ether contracts. This performance highlights a growing trend of algorithmic dominance in venues traditionally reserved for crowd-sourced forecasting.

Key details

The bot identified fleeting moments where the combined price of "Yes" and "No" contracts summed to less than $1. In a functional market, these two outcomes always settle at a total of $1. When the combined price dipped to $0.97, the bot bought both sides to lock in a guaranteed three-cent spread. Each trade generated an average of $16.80 in profit.

Polymarket serves as the primary venue for this activity. The company currently holds a $9 billion valuation following a $2 billion investment from Intercontinental Exchange (ICE), the owner of the New York Stock Exchange.

| Metric | Value | | :--| :--| | Total Profit | ~$150,000 | | Number of Executions | 8,894 | | Edge per Trade | 1.5% to 3% | | Typical Trade Size | ~$1,000 | | Market Depth (per side) | $5,000 to $15,000 | | Venue Valuation | $9 billion |

CEO Shayne Coplan leads the Manhattan-based firm, which recently added Donald Trump Jr. as an advisor. While the platform operates globally in over 160 countries, it maintains restrictions in 33 jurisdictions, including the United Kingdom, France, and Germany. In the United States, the platform re-entered the market in late 2025 as a regulated Designated Contract Market (DCM) under Commodity Futures Trading Commission (CFTC) oversight.

Market context

Liquidity in five-minute prediction markets remains thin compared to traditional crypto exchanges. Active sessions on Polymarket show order-book depth between $5,000 and $15,000 per side. This is several orders of magnitude smaller than Bitcoin perpetual swap books on Binance or Bybit. This thin liquidity creates a ceiling for automated strategies. A trader attempting to deploy $100,000 per trade would immediately erase the spread and eliminate the arbitrage opportunity. The game belongs to retail-sized accounts trading in the low four figures.

Business model changes are also altering the profitability of these bots. In early 2026, the platform introduced taker fees for high-frequency markets. Fees for five-minute crypto contracts now reach up to 1.56% at 50% probability. These fees fund a maker rebate program, where liquidity providers receive 20% to 25% of collected fees. This shift makes the 1.5% to 3% arbitrage edge much tighter for aggressive bots that cross the spread.

Prediction market angle

The market is increasingly pricing these contracts based on derivatives data rather than independent crowd sentiment. As AI systems arbitrage prediction markets against external options and perpetuals, the prices begin to mirror broader crypto markets. This transition threatens the role of prediction markets as unique sources of probability. If a bot can see a price move on Binance before it reflects on a five-minute Polymarket contract, the prediction market becomes a lagging indicator rather than a forecasting tool.

What this means

For retail traders, these glitches represent a diminishing window of opportunity. The introduction of taker fees and the entry of more sophisticated algorithms mean that manual traders cannot compete on speed. Institutional desks are restricted by the lack of depth, but mid-sized automated systems are filling the gap.

Regulatory status continues to be a factor for growth. Federal approval from the CFTC allows for nationwide operations in the US, but certain states like Nevada and Tennessee continue to challenge the platform in court. These legal disputes focus on whether event contracts constitute unlicensed gambling. The outcome of these cases will determine if the platform can sustain its $3 billion monthly volume reached in late 2025.

Follow FlowFrame for real-time data on prediction market spreads and liquidity shifts.