By 2025, prediction markets had fully entered the mainstream. Annual trading volume quadrupled, and control of the market became concentrated among a few major platforms. This is according to a new report by CertiK, a company specializing in blockchain security.
According to the study, total trading volume increased from $15.8 billion in 2024 to $63.5 billion in 2025. Activity did not decline after the end of the U.S. election cycle and continued through January 2026.
Monthly trading volume on prediction markets in 2025. Source: CertiK
This resilience is important. It shows that trading around the elections was not a one-off burst of interest. Rather, it became an entry point for new users who continued to trade actively after the political race ended.
Notably, the week ending January 18 set a record with a nominal volume of about $6 billion. This reflects how quickly prediction markets transformed from a niche crypto product into a high-volume, institutional-scale platform.
However, the main takeaway from CertiK is different. The next stage of growth is limited not so much by smart contract hacking risks, but by higher-level issues. This refers to onboarding mechanisms, what actually underlies the volume figures, and the systems that determine who gets payouts and under what conditions.
A Market of Three Platforms and the Risk of a Single Point of Failure
According to CertiK, more than 95% of global prediction market volume now comes from just three platforms. Each has chosen its own strategy for dominance.
Kalshi operates as a regulated platform in the U.S. and focuses on a compliance-first model. Polymarket holds the largest share among crypto-native and international audiences.
Meanwhile, Opinion has become the fastest-growing player. The platform uses ecosystem incentives and grew from almost zero to about 30% market share in just a few months.
Such concentration turns any operational problem into a systemic one.
A failure at one of the key platforms can no longer be considered a local incident. It is a risk for the entire market, potentially affecting liquidity, data sources, and user balances. Especially since brokers and mass distribution channels are beginning to treat prediction market probabilities as a new type of information product.
See also: Bitcoin Has Not Bottomed Yet. CryptoQuant Urges Patience
The report cites a December 2025 incident with Magic.link, a third-party authentication provider for Polymarket. According to CertiK, this is a clear example of the sector’s most vulnerable point.
Accounts using Web2 logins via email or social networks were compromised. Funds in such accounts were at risk, although the on-chain settlement layer remained protected.
As CertiKemphasizes, this was an identification failure, not a settlement failure. The incident demonstrated the downside of the Web2.5 onboarding model: a more convenient user experience in exchange for centralized points of failure.
For an industry that builds its marketing around decentralization, this is an inconvenient conclusion.
Prediction markets can provide fully collateralized on-chain settlements, but still retain the same third-party risks as classic fintech. This refers to authentication, account recovery, and access control at the platform level.
When Volumes Deceive, but Probabilities Still Tell the Truth
The report also draws an important distinction between two concepts that are often conflated in the crypto industry. The first is trading volume as an indicator of adoption. The second is probabilities as a source of information.
CertiK notes that incentive programs can inflate activity, but this does not mean the quality of forecasts improves.
According to the study, wash trading remains widespread. During peak airdrop farming periods, artificial volume on some platforms reached 60%. Such figures can create the illusion of deep liquidity and high organic participation, especially for potential institutional investors.
However, according to CertiK, a more important question is whether probabilities retain their value even if the trading tape is distorted.
See also: Bitcoin Funding Turns Negative. Is the Market Preparing for a Reversal or Another Drop?
The report states that wash trading has indeed inflated volume figures, but so far has not affected price accuracy. The resulting probabilities remain useful for forecasting.
This is where the tension arises for platforms that want to reach the level of traditional finance. Even if some activity is driven by incentives, platforms can position themselves as information services if their probabilities remain reliable.
But this presents a more difficult choice for market leaders.
If further growth depends on trust and information quality, platforms will have to take a tougher stance on practices that temporarily increase volumes but undermine the reputation and transparency needed to attract institutional capital.
Migration Between Networks and New Execution Infrastructure
Behind the impressive growth numbers, CertiK sees a deeper shift in how liquidity is actually executed in prediction markets.
Polygon, according to the report, maintained historical dominance up until the November election cycle. However, from late 2025, volumes on BNB Chain began to rise sharply. This spike coincided with the aggressive launch of incentives by Opinion.
By the week ending January 19, activity on BNB Chain had effectively changed the previous hierarchy. The network captured the largest share of weekly flows, and off-chain settlements took a back seat. This happened even amid record Kalshi performance during NFL playoff trading.
Prediction market volume by blockchain. Source: CertiK
This shift is important not only as a competition between ecosystems. It changes the very architecture of the market. The choice of network determines who can participate, how settlements are made, and what trading models become possible.
CertiK notes that many on-chain platforms are gradually moving away from automated market makers and switching to central limit order books deployed directly in high-performance networks. This model provides tighter spreads and familiar mechanics for professional traders.
In practice, this makes prediction markets increasingly similar to classic exchanges. Along with this come related risks: front-running and vulnerabilities related to the order of transaction inclusion in a block, typical for public networks and MEV mechanics.
The Oracle Problem and the Moment When ‘Truth’ Becomes a Payout
If there is one systemic risk that unites the entire growth of prediction markets, it is the resolution stage. It is at this moment that probabilities turn into real payouts.
Prediction market security risks. Source: CertiK
CertiK calls oracle manipulation the main technical attack vector, since resolution mechanisms directly control fund distribution.
In addition, in 2025, all major platforms already experienced disputes due to vague market formulations. This happened especially often in political events, where official results could be disputed or interpreted differently.
The report details the resolution models of the leading players.
Polymarket uses an optimistic oracle from UMA. The result is confirmed automatically if no objection is filed within the set period. In case of a dispute, the decision is put to a vote by UMA token holders.
Kalshi relies on centralized arbitration. Outcomes are determined by appointed arbitrators, who rely on authoritative sources.
Opinion uses consensus oracles. The outcome must be confirmed by designated participants who reach a joint decision.
See also: USDC Activity Hits Record High Amid Market Correction
Each model is based on its own trust assumptions. Optimistic oracles work quickly for clear outcomes but are vulnerable in borderline cases. For example, large token holders can influence voting in low-liquidity disputes.
Centralized arbitration is more predictable but requires trust in the platform operator. The consensus model distributes authority but still depends on incentives and the good faith of appointed resolvers.
As the market grows, it becomes increasingly difficult to ignore these trade-offs.
When prediction markets remain a niche crypto product, isolated controversial cases can be survived. But if their probabilities begin to appear in mass channels or are used by institutional players for risk assessment, any dispute over resolution becomes not just a local incident but a crisis of trust and governance.
