Huobi Growth Academy | Predictive Market Depth Research Report: Liquidity Paradigm, Industrial Leap, and New Semantic Revolution

I. Historical Evolution and Industry Landscape of Prediction Markets

Prediction markets, as a mechanism for pricing future events, have evolved over more than thirty years from academic experiments and gray-area gambling to an independent asset class that combines informational value, liquidity scale, and financial attributes. These markets are centered around the structure of "price equals probability," using real funds to reflect market participants' aggregated judgments on the probability of a certain event occurring: a binary contract that settles at $1 or $0 has a trading price between $0 and $1, directly presenting market consensus. For example, when the price of a certain event contract is $0.62, it means "the market believes the probability of the event occurring is about 62%." This mechanism, based on the aggregation of views from dispersed participants, essentially constructs a quantifiable, verifiable, and real-time updated public good of information, which is not only different from recreational gambling but also distinct from the bookmaker structure of binary options. Instead, it is a hybrid information financial infrastructure that combines market efficiency, collective wisdom, and dynamic trading capabilities. Unlike the zero-sum mechanism of gambling, the overall structure of prediction markets presents a "positive-sum information output": the platform charges a small fee, while the core value comes from the probability signals aggregated by the market. These signals can be cited by the media, modeled by research institutions, used by companies for risk management, and can also be directly embedded as pricing nodes in other financial derivatives and Web3 protocols, possessing strong externalities and social value.

The foundation of modern prediction markets can be traced back to the Iowa Electronic Markets (IEM) in 1988. This early experiment, led by academic institutions, allowed participants to trade contracts representing the winning probabilities or vote shares of candidates with small amounts of money, with a clear goal of improving prediction accuracy. Numerous studies have shown that between 1988 and 2004, the IEM significantly outperformed most traditional polls in predicting U.S. elections, with its probability signals demonstrating an ability to reflect real trends earlier.

What truly propelled the industrial leap of prediction markets was the emergence of a new generation of platforms after 2020, underpinned by the maturity of Layer 2, stablecoins, and cross-chain infrastructure, represented by the "dual oligopoly" formed by Polymarket and Kalshi in 2024-2025. Polymarket represents the comprehensive maturity of the decentralized route: based on Polygon and multi-chain expansion, it has achieved a product form that combines experience and censorship resistance through a Central Limit Order Book (CLOB), low-friction deposits, gas-free trading, and UMA's optimistic oracle. During the 2024 U.S. election, its monthly trading volume reached $2.6 billion, with annual cumulative trading exceeding $10 billion. Its significant dissemination effect on media and social networks has constructed a flywheel of "opinion → position → dissemination," making it the preferred platform for Web3 users to enter prediction markets. Even after being penalized by the CFTC, it restructured its U.S. market presence by acquiring the licensed exchange QCEX, further indicating that compliance has become the core direction for the development of the sector. In parallel, Kalshi represents a completely different path: compliance, regulatory certainty, and penetration into mainstream financial channels. Kalshi obtained CFTC Designated Contract Market (DCM) status in 2021 and subsequently secured a clearing license (DCO), becoming a federally compliant event contract exchange in the U.S. Its centralized matching structure is closer to traditional exchanges, supporting deposits in USD and USDC, and directly providing event contracts through partnerships with brokerages like Robinhood on mainstream investor interfaces. After the explosion of sports and macroeconomic contracts in 2025, Kalshi's weekly trading volume once reached $800-900 million, capturing a market share of 55-60%, effectively becoming the infrastructure for domestic prediction markets in the U.S. Unlike Polymarket's on-chain openness, Kalshi's advantages lie in the institutional participation, brand trust, and traditional channel distribution capabilities brought by compliance certainty. Together, they form a dual core of "on-chain composability" and "compliance usability."

Beyond the dual oligopoly, new platforms and vertical tracks are rapidly emerging, further expanding market boundaries. Opinion leveraged BSC ecosystem traffic and airdrop incentives to surpass hundreds of millions in scale within its first week of launch; Limitless meets crypto traders' demand for volatility products through short-cycle price predictions within the Base ecosystem; PMX Trade in the Solana ecosystem directly tokenizes Yes/No contracts, exploring the deep integration of prediction markets and DEX liquidity. Sports-related platforms like SX Network, BetDEX, and Frontrunner have become the largest vertical scenes due to their high frequency and stickiness, while "creator economy prediction markets" represented by Kash, Melee, and XO Market directly financialize opinions, turning KOL viewpoints into tradable assets. Meanwhile, TG Bot and aggregator toolchains represented by Flipr, Polycule, and okbet are becoming another rapidly developing direction, compressing complex prediction interactions into chat interfaces, providing cross-platform price tracking, arbitrage, and fund flow monitoring, forming a new ecosystem of prediction markets akin to "1inch + Meme Bot."

Overall, prediction markets have gradually completed three leaps in their thirty-year evolution: from academic experiments to commercial gambling exchanges, then from on-chain experiments to dual-core platforms of compliance and scaling, ultimately differentiating into highly diversified forms in vertical scenes such as sports, crypto markets, and the creator economy. The window for general-purpose platforms is narrowing, while true incremental growth is more likely to come from deeply verticalized scenes, the data and tool layers surrounding the ecosystem, and the degree of integration of prediction market signals with other financial systems. Prediction markets are accelerating from a "gray-area toy market" toward becoming "an important infrastructure of the global information and financial system."

II. Structural Challenges of Prediction Markets

After more than thirty years of iteration, prediction markets have transitioned from experimental products to a financial-grade infrastructure stage with gradual participation from global users and institutions. However, their development still faces three major structural bottlenecks that cannot be circumvented: regulation, liquidity, and oracle governance. These three are not independent; they are interconnected and mutually restrictive, determining whether prediction markets can grow from "gray innovation" into a "compliant and transparent global information and derivatives system." Regulatory uncertainty limits institutional capital entry, insufficient liquidity weakens the effectiveness of probability signals, and if oracle governance cannot provide a reliable adjudication mechanism, the entire system may fall into a quagmire of manipulation and disputes over outcomes, failing to truly become a trusted source of information for the external world.

Regulatory issues are the primary bottleneck for prediction markets, with their complexity particularly prominent in the United States. Whether prediction markets are classified as commodity derivatives, gambling, or a type of securities investment contract corresponds to different regulatory paths. If viewed as commodities and derivatives, they fall under CFTC regulation and are treated similarly to futures exchanges, requiring the application for DCM (Designated Contract Market) and DCO (Clearing Organization) licenses, which have high thresholds and costs. However, if successful, they gain legal status at the federal level, as exemplified by Kalshi. If classified as gambling, they must apply for gambling licenses in all 50 states, leading to exponentially rising compliance costs that nearly block the possibility of a nationwide platform. If considered securities, they trigger strict SEC regulation, posing significant potential risks for DeFi prediction protocols with token designs or revenue promises. The fragmented and overlapping U.S. regulatory system places prediction markets in a repeatedly disputed gray area. For example, the lawsuit between Kalshi and the New York Gaming Commission centers on whether the CFTC has exclusive regulatory authority over event contracts. This ruling not only affects whether Kalshi can operate smoothly nationwide but also concerns the institutional trajectory of U.S. prediction markets for the next decade. Furthermore, the CFTC's enforcement actions against Polymarket and its classification of Crypto.com's sports event contracts indicate that regardless of whether a platform's shell is "decentralized," as long as it provides a front end to U.S. users and facilitates transactions, it will essentially be viewed as an unregistered compliance activity based on derivatives or binary options, incurring corresponding legal responsibilities.

Outside the U.S., various global jurisdictions generally continue the "binary framework": either incorporating prediction markets into gambling regulatory systems or into financial derivatives systems, with very few new laws specifically for prediction markets. Countries like the UK and France maintain an open attitude towards event betting under online gambling regulations, but regulatory tools such as geographic blocking, payment bans, and ISP blocking make it difficult for prediction market platforms to reach mainstream users before obtaining licenses. For entrepreneurs, the "technological neutrality" defense can no longer evade legal risks; offshore companies, DAOs, or decentralized front ends cannot ensure immunity from regulation. The only paths for long-term survival are threefold: either embrace licensing head-on like Kalshi; or maintain complete offshore and fully open-source decentralization while accepting the cost of absence from the mainstream market; or pivot to building compliant infrastructure, providing technical services (KYC, risk control, prediction data API, etc.) for licensed institutions. Regulatory uncertainty limits institutional capital participation and restricts the depth of connections with traditional finance, making it difficult for prediction markets to truly scale.

III. Value Innovation and Future Opportunities in Prediction Markets

After several rounds of reshuffling due to the constraints of regulation, liquidity, and oracle governance, truly valuable innovations in prediction markets are beginning to shift from "single platform competition" to the "primitive layer" and "infrastructure layer." Simply put, what has been done in the past decade is "a new prediction market website"; while in the next decade, the incremental growth is more likely to come from "abstracting event contracts into informational derivatives and embedding them into the entire DeFi and financial system," transforming prediction markets from an application into a piece of DeFi Lego that can be pieced together. The binary contracts of events themselves are just the starting point; once contracts become standardized, composable, and collateralizable asset units, a complete set of derivative layers such as perpetuals, options, indices, structured products, lending, and leverage can naturally grow around them. The "event market" explored in D8X, Aura, and parts of dYdX v4 design essentially projects "whether it occurs" into a price space of 0-1, further allowing high-leverage trading, enabling traders not only to bet on event directions but also to trade volatility and sentiment. Protocols like Gondor allow users to collateralize Polymarket's YES/NO shares to borrow stablecoins, transforming originally static locked long-term event positions into reusable collateral assets. The protocol then dynamically adjusts LTV and liquidation logic based on market probabilities, financializing "opinions" into reusable capital tools. Further up are index and structured products similar to PolyIndex, which bundle a basket of events into ERC-20 index tokens, allowing users to gain comprehensive exposure to a certain theme with one click, such as the "U.S. Macro Policy Uncertainty Index" or "AI Regulation and Subsidy Implementation Event Basket." In the context of asset management, prediction markets are no longer isolated markets but become a new asset class that can be incorporated into portfolio configurations by asset managers.

The truly mid-to-long-term valuable "shovel opportunities" concentrate on four levels. The first is the truth and rules layer, namely the new generation of oracles and arbitration protocols. How to avoid the reoccurrence of disputes like UMA in terms of economic incentives and governance structures, and how to use standardized, modular tools to help ordinary users create "clearly defined and arbitrable" event markets will directly determine the extent to which prediction markets can be trusted by institutions and public sectors. The second is the liquidity and capital efficiency layer, where AMMs, unified liquidity pools, collateralized lending, and yield aggregation protocols tailored for prediction markets can transform dormant event positions into reusable assets, bringing new asset classes to DeFi and providing platforms with a thicker economic moat. The third is the distribution and interaction layer, including social embedded SDKs/APIs, media one-click access components, professional terminals, and strategy tools. These directions determine the "entry forms" of prediction markets and decide who can stand at the intersection of information and trading to earn continuous fees and technical service fees. The fourth is the compliance technology and security layer, which helps licensed institutions safely access prediction market data within the regulatory framework through refined geographic fencing, KYC/AML, risk monitoring, and automatic reporting across multiple jurisdictions, allowing event prices to truly enter the asset management, investment research, and risk management processes. Finally, the rise of AI provides a new closed loop for binding prediction markets with capital markets. On one hand, AI models can act as "super traders" in prediction markets, trading with stronger information processing and pattern recognition capabilities, thereby improving market pricing efficiency; on the other hand, prediction markets can serve as "real-world scoring grounds" for AI capabilities, quantitatively assessing model quality through real profits and long-term calibration metrics, providing an external, hard-constraint evaluation system for "AI research reports, AI investment advisors, and AI strategies." For investors, projects that understand derivative design, can safely utilize event prices within regulatory boundaries, and bridge AI with traditional finance are likely to grow into key infrastructure assets in the entire "informational derivatives" sector in the next cycle.

IV. Conclusion

From the betting markets of the 16th-century papal elections to the predictions of presidents on Wall Street in the 20th century, and then to IEM, Betfair, Polymarket, and Kalshi, the evolution of prediction markets is essentially a history of humanity's attempts to approach "more accurate probabilities" through systems and incentives. Today, as mainstream media trust continues to decline and social platform signals are mixed with noise, prediction markets materialize the "cost of saying the wrong thing" through prices, compressing scattered information and judgments from around the world into a quantifiable, verifiable probability curve. It is not a perfect truth machine, but it provides a more verifiable public signal than slogans and emotions. Looking ahead, the ultimate fate of prediction markets may not be the emergence of a single platform larger than Polymarket, but rather becoming an "information and opinion interaction layer" embedded in social media, news websites, financial terminals, games, and creator tools; ubiquitous like a "like button," allowing every opinion to naturally correspond to a tradable probability; continuously producing incentivized "collective predictions" in the game involving both humans and AI, feeding back into decision-making and governance. To truly reach that point, the sector must first cross three thresholds: the regulatory threshold, the liquidity threshold, and the oracle governance threshold. These three thresholds are the stage for the next generation of infrastructure and emerging primitives. For entrepreneurs and investors, prediction markets are by no means a sector that has already been "completed"; on the contrary, it has just completed the first phase from concept to industrial prototype. What will truly determine whether it can become "Web3-level information infrastructure" is the ongoing innovation and institutional integration around rules, liquidity, and oracle governance in the next 5-10 years. In this information war worth billions of dollars, the winners are often not the loudest voices but those builders who quietly solidify the "shovels" and "roads."