After a decade of honing, is the next major player about to take the stage in the prediction track?

The evolution of crypto prediction markets is very interesting because they were once considered a “debunked” track, and it took ten years to achieve PMF (product-market fit), with developments that exceeded market expectations. Sometimes, in the crypto space, it’s not always appropriate to draw conclusions too early.

The concept of prediction markets itself isn’t new; it has existed in crypto for a long time. In 2015, the Gnosis project began development; in 2018, Augur officially launched as a decentralized prediction market platform based on Ethereum, allowing users to create and predict future events and settle using cryptocurrency.

In 2020, Polymarket (based on Polygon) was also launched, but it remained marginalized. Coupled with regulatory factors, it struggled to operate. In its early days, Polymarket’s monthly trading volume was only a few million dollars; Augur’s TVL plummeted nearly 80% after the 2020 election, dropping from its peak to a few million dollars. The whole industry’s TVL peaked at around $7 million, with monthly trading volume below $100 million. Regulatory pressures (like the CFTC viewing it as “gambling”) and underdeveloped oracles (prone to manipulation) further suppressed growth.

The real explosion of prediction markets didn’t happen until 2024. Especially, the 2024 US election became a turning point. Polymarket’s election prediction market trading volume exceeded $2.7 billion, with total monthly trading volume on the platform soaring from $62 million in May to $2.1 billion in October, over 30 times growth. The annual notional trading volume reached $16.3 billion, far surpassing previous totals.

Why did it take ten years to achieve PMF?

First, there were technical and user experience barriers in the early crypto space. The concept of prediction markets was good and seemed to have great demand, but when it came to actual user experience, it excluded almost all users. For example, early Augur was built on Ethereum L1, with extremely high transaction costs and painfully slow confirmation times. Ordinary users also had to master wallets and complex user interfaces, which imposed a high learning curve. These high barriers led to insufficient liquidity and worries about manipulation.

Second, regulatory pressure was always present. The US CFTC (Commodity Futures Trading Commission) classified prediction markets as “gambling” or derivatives and increased scrutiny after 2018. During this period, Augur was fined for bets on sensitive events; in 2022, Polymarket paid a $1.4 million fine and exited the US, with its founder Shayne Coplan (born 1998) having his New York apartment raided by the FBI, who seized his electronics (though he was not arrested). The regulatory gray area prevented institutional funds from entering. Regulatory pressure made it difficult for liquidity to increase.

Third, market narrative shifts. In the 2016-2018 crypto era, most users were more focused on speculation than practical tools; from 2020-2023, the DeFi/NFT boom distracted attention, and prediction market TVL hovered at only $7 million. Without mainstream event drivers, it was hard to accumulate liquidity.

Fourth, oracles were immature and easy to manipulate.

Yet 2024 was a turning point. As mentioned above, the 2024 US election was a catalyst, but it was not the only factor.

From 2024 onwards, prediction markets truly took off. In addition to Polymarket, centralized prediction platform Kalshi also emerged. By 2025, prediction market trading volume reached $27.9 billion (up 210% year-on-year), with weekly peaks of $2.3 billion. The combined TVL of Polymarket and Kalshi exceeded $20 billion. Both platforms reached valuations in the tens of billions. Prediction markets suddenly became the market’s darling.

So, what are the driving factors?

Contrary to the obstacles faced between 2015-2024, these barriers have gradually been removed, leading to qualitative improvements in user experience and other aspects.

First, changes in technology barriers and user experience. Polygon and Base L2 networks reduced gas fees to mere cents, and transaction speeds increased tenfold. Platforms like Polymarket optimized their UI and supported one-click betting with stablecoins, attracting non-crypto natives. In addition, DeFi experienced huge growth, providing deep liquidity. For users, participating in prediction markets has become very convenient. Kalshi is a centralized prediction platform integrated with apps like Robinhood, making participation even easier.

Second, changes in regulation. After the 2024 US election, regulators promoted crypto-friendly policies. In 2025, the CFTC approved regulated platforms like Kalshi. The SEC/CFTC clarified the legality of “spot commodity crypto,” and stablecoin legislation passed Congress. Overseas, Switzerland still had a blacklist, but the overall environment shifted from hostile to supportive, with institutional funds pouring in (e.g., ICE invested $2 billion).

Third, changes in market narrative. In this cycle, there hasn’t been a particularly strong narrative. Instead, real utility has become the market’s focus. Catalyzed by the 2024 election predictions, Polymarket expanded into sports, economics, technology, etc., aided by media coverage (e.g., CNN/Bloomberg) and social network promotion, fueling the boom in prediction markets.

Fourth, both institutions and communities are driving growth. a16z actively participated, introducing the narrative of “event-driven financial infrastructure,” while community users also actively engaged, pushing TVL higher.

Fifth, prediction markets have gradually evolved from “gambling” into a new type of signaling mechanism, similar to providing real-time probability signals.

From the ten-year evolution of prediction markets, an interesting conclusion emerges: not all “debunked” tracks necessarily lack PMF; sometimes the conditions just aren’t mature yet. In crypto, this phenomenon is especially evident, as the foundational infrastructure was underdeveloped in the first decade (expensive/slow/poor user experience, etc.), making it hard for many attempts to reach ordinary users. Perhaps in the future, tracks like Crypto Gaming, social, AI agent, DePIN, digital identity, etc., still have opportunities to shine, even if some may seem finished for now.