Prediction markets have grown rapidly. Platforms like Polymarket and Kalshi have evolved into multi-billion-dollar businesses, and traditional exchanges – Eurex, Euronext, CME Group, CBOE, ICE and Nasdaq – are increasingly active in the space. For example, these exchanges are using prediction markets data to add analytical value, generating real-time signals on market expectations for political, economic or social outcomes.
In ESMA's latest Trends, Risks and Vulnerabilities report, published on 10 September 2026, a section dedicated to prediction markets has been included for the first time. ESMA highlights a range of investor protection and market integrity concerns, and draws attention to the widening gap between Europe and US's regulatory approaches.
But why Europe is falling behind? The answer: regulatory complexity. A clearer regulatory landscape for prediction markets in Europe might be the path to manage the risks that are flagged.
The risks
ESMA identifies a range of significant risks:
- Retail investor harm. Prediction platforms offer "speculative gambling environments" without the investor protections associated with regulated financial products. Their gamified structure, emotional dynamics and social media-driven promotion expose inexperienced retail investors to "significant risks of financial loss, addictive behaviour and exploitation by more sophisticated traders." The data backs this up: a Wall Street Journal analysis found 67% of profits on Polymarket accrued to just 0.1% of accounts; Bloomberg reported that most users lose money.
- Insider trading. ESMA observes that prediction markets are "rife with insider trading." The report cites several recent incidents, such as the US soldier who was criminally charged for using classified information to bet on the capture of Nicolás Maduro. In April 2026, suspected tampering with weather sensors used to settle Polymarket contracts prompted Météo-France to file a police complaint.
- Contract resolution risk. Prediction markets depend on clearly defined event outcomes and high-quality data sources for settlement. Ambiguous event definitions, manipulated reference data, opaque resolution mechanisms and settlement failures, particularly in decentralised settings (see below), all create risks that may not be immediately apparent to participants.
- DeFi and AI amplifiers. Decentralised prediction markets operate without identifiable intermediaries or central governance, limiting accountability. Smart contract transactions are irreversible, and reliance on external data feeds and oracles introduces manipulation risk. Meanwhile, AI-generated trading signals and automated bots may exacerbate information asymmetries and facilitate coordinated trading behaviour. AI-generated misinformation and coordinated online campaigns could further distort pricing dynamics.
The regulatory framework
The report touches briefly on the regulatory treatment of prediction markets, but the complexity in this space merits closer attention.
In the EU, the classification of event contracts is fragmented. Depending on their characteristics, they may qualify as financial instruments under MiFID II, fall within scope of MiCA where they are DLT-based (and do not qualify as financial instruments), or be treated as gambling products under national law. In each case, EU authorisation is required, and the largest platforms currently hold none. Where event contracts do qualify as financial instruments, they would generally be classified as derivatives and caught by national product intervention measures on binary options, meaning they are banned for retail.
The practical result: prediction markets have not gained significant traction in the EU compared with the US.
A broader question arises from this framework. The regulatory perimeter may present a significant constraint on the ability to address the very risks the report identifies. For example, the Market Abuse Regulation can only help where prediction market contracts fall within the financial regulatory perimeter. The effectiveness of enforcement against manipulation, insider trading and consumer harm will depend on which regulatory framework applies to a given contract, and that classification exercise is not always straightforward or consistent across Member States.
The geographic access restrictions applied by Polymarket and Kalshi also raise questions. Both platforms state that users in "some, but not all" EU countries are prohibited from placing orders. ESMA notes that it is "unclear why all EU Member States are not included in the list of restricted jurisdictions" and critically, that such restrictions do not prevent EU users accessing the platforms through VPNs. The platforms say VPN use is prohibited and detectable, but ESMA is not certain on the "practical effectiveness of these restrictions".
What to watch
It does seem the US-based platforms are looking to work with regulators to expand into the EU. Clearer guidance on how to do so would help growth in this space whilst managing risk.
Malta is the first EU Member State to publicly explore a dedicated regulatory framework for prediction markets, describing the sector in March 2026 as one of "rapid global momentum" offering "significant opportunities for innovation."
Whilst we wait for regulatory development and clarity, ESMA's current views are clear: the prediction market space will be closely monitored as it develops rapidly and new risks emerge.
