Two companies dominate 97%, trading volume surges by 1100%: Reshaping the landscape of prediction markets and the next round of entrepreneurial opportunities

2026-03-05 12:23:34

Collection

The global forecast market's total trading volume is expected to be approximately $50.25 billion in 2025. If maturity is defined by trading volume rather than narrative, the forecast market will truly complete its transition from "event-driven short-term curiosity" to "sustained market" in 2025.

# Introduction: Redefining the Boundaries of Prediction Markets

Prediction markets are markets that allow participants to trade on the outcomes of uncertain future events, with contract prices reflecting the market's consensus judgment on the probability of those events occurring. They have demonstrated significantly greater accuracy than expert predictions and polls in areas such as political elections, macroeconomics, sports events, crypto assets, and corporate events.

The essence of prediction markets is as a tool for "financializing information"—price equals probability. In fields characterized by high uncertainty and strong subjective judgment, prediction markets have significant advantages.

The total trading volume of global prediction markets is expected to be around $50.25 billion in 2025. If we define maturity by trading volume rather than narrative, prediction markets will only truly complete their leap from "event-driven short-term curiosity" to "sustained markets" in 2025.

Kalshi has validated that the industry is not just about "having trading volume," but is beginning to demonstrate commercial capabilities—reporting approximately $260 million in fee revenue. Nevertheless, prediction markets have not yet truly entered a growth phase; compared to the global futures market in mature finance, which has annual trading volumes in the trillions, it resembles financial futures in 1982 rather than cryptocurrencies in 2020.

Unlike most financial innovations, prediction markets have not experienced long-tail competition but have rapidly consolidated into two platforms: Kalshi and Polymarket, which together account for over 97.5% of market share, while all other platforms combined have only about $1.25 billion in trading volume, belonging to a marginal ecosystem.

## I. The Essence of Prediction Markets: An Information Production Mechanism in a Non-Attention Economy

Prediction markets are no longer simply an innovation in trading forms but are evolving into an information production mechanism in a non-attention economy.

The core differences from traditional attention economies are:

Value does not depend on clicks, traffic, or popularity.
The core assets are the quality of cognition and information.
Market participants seek verifiable, tradable, and citable judgments rather than short-term attention exposure.

Under this logic, the competitive landscape for prediction markets has also shifted:

Brokerage research systems.
Consulting firm judgment systems.
Media narrative authority.
Probability outputs from AI training.

In other words, this is a market for pricing future cognition.

The true watershed moment for the industry at this stage lies not in technology but in three aspects: whether it can form sustained information liquidity; whether it enters a "weakly regulated tolerable zone" rather than a gray arbitrage zone; and whether it is treated by institutions as a decision input rather than a retail entertainment tool. Once these three points are established, the form of prediction markets will be closer to a hybrid of Bloomberg + exchanges + polling agencies rather than Web3 applications.

Problem Definition Authority: The Severely Underestimated Core Asset

The vast majority underestimate the most core asset of prediction markets—not liquidity, but the ability to define problems.

Whoever controls problem definition controls: the information entry point, trading context, and probability interpretation rights. This is highly similar to the power structure of index companies (such as MSCI). A well-designed market question is essentially a tradable cognitive framework.

## II. Why is the Value of Prediction Markets Suddenly Reassessed in the 2024-2026 Cycle?

2025 is not coincidentally becoming a turning point; it is the result of the overlap of three types of structural factors.

### 2.1 Clarification of Regulatory Expectations

In 2024, the regulatory attitudes of multiple states in the U.S. and the CFTC towards event contracts are becoming clearer.
Kalshi's legitimate path opens the door for traditional institutional funds, leading to a sudden increase in institutional trading volume.
Traditional investors are beginning to view prediction markets as "event trading tools that can contribute to alpha," rather than gray gambling.

### 2.2 Concentration of Trading Scale + Continuous Event Supply

In the past, prediction market events were often concentrated in politics or single events, with short trading cycles and high volatility.
In 2025, high-frequency events (sports, corporate KPIs, crypto market events) emerge, allowing the market to continuously absorb funds.
Continuous events create a self-reinforcing cycle of liquidity: liquidity brings information depth → attracts more trading → price signals become more accurate.

### 2.3 Marginal Amplification of Information Demand

Although data is flooding in the AI era, "probability credibility" has become a scarce asset.
Quant funds, hedge funds, and corporate decision-making departments are starting to view prediction market prices as a source of real signals.

The core logic is not user growth in traffic but the concentration of liquidity triggered by capital and information demand—this is the true turning point for prediction markets.

### 2.4 Three Structural Forces Overlapping

Force One: The "Marginal Failure" of Traditional Research Systems is Becoming Apparent

Over the past decade, sell-side research has significantly lagged in predicting macro turning points; buy-side firms are gradually viewing "the speed of consensus formation" as a source of alpha; expert models are increasingly resembling narrative engineering rather than probability discovery.

Prediction markets provide a different paradigm: it is not about "who is smarter," but "who is willing to pay for a judgment." The exposure of funds itself becomes an information filter.

Force Two: After the Rise of AI, Human Society Needs "Real Signal Sources" More Than Ever

Large models can generate judgments but cannot bear risks. The uniqueness of prediction markets lies in their irreplaceable mechanism advantages:

|-------|----|------| | Mechanism | AI | Prediction Markets | | Output Judgment | | | | Bear Loss | | | | Prevent Nonsense | | | | Information Self-Correction | | |

Thus, it becomes one of the few systems in the AI era with a fact-anchoring mechanism, which is why more and more quant funds are starting to treat prediction market prices as exogenous variables.

Force Three: Web3 Solved a Key Constraint—Settlement Credibility

The biggest problem for early prediction markets was not the lack of predictions but: who would be the market maker? How to prevent defaults? How to enable global participation? On-chain settlement reduces trust from "trust the operator" to "trust code execution," giving prediction markets the ability to expand across jurisdictions for the first time.

## III. Comparison of Leading Platforms by Scale (Actual Size in 2025)

### ① Kalshi --- Current Liquidity Center

The nominal trading volume in 2025 is expected to be about $23.8 billion, with a year-on-year growth of over 1100%.
At one point, it accounted for 55%-60% of the industry's weekly trading volume, becoming the most liquid market.
During certain statistical periods, its global market share rose to 62.2%.
Monthly trading volume once reached the level of $1.3 billion.
The growth momentum mainly comes from the compliance path opening traditional fund entry, rather than the expansion of crypto users.

Kalshi has chosen a completely different strategy: actively entering the regulatory framework and defining prediction markets as "event contract exchanges," attempting to replicate the legitimacy path of the futures market. Short-term growth may be slow, but if successful, it will open the floodgates for pension/RIA/institutional fund allocation.

### ② Polymarket --- Crypto-Native Liquidity Hub

The total trading volume for 2025 is expected to be around $22 billion.
In certain months, it still maintains a monthly trading scale of hundreds of millions.

Polymarket has taken a global permissionless liquidity path: rapidly forming event coverage density, using on-chain technology to reduce participation friction, and replacing compliance depth with trading activity.

Its true value lies not in trading volume but in establishing the world's first "real-time political probability curve"—such data has never existed in traditional systems.

### ③ Second-Tier Platforms (Total Share is Extremely Small but Represents Future Differentiation Directions)

Despite the high concentration of the market, several exploratory platforms have emerged, such as Azuro and TrendleFi. These platforms combined contribute only about $1.25 billion in trading volume, indicating that the industry has not yet entered a "hundred flowers blooming" phase but is still in the stage of infrastructure rights confirmation.

Augur represents the limitations of the first generation of decentralized experiments: overemphasizing "trustlessness" while neglecting the real trader experience, failing to solve problem distribution and liquidity acquisition. This indicates that prediction markets are not purely a technical issue but a market design issue.

|------------|-----------|--------------|-----------| | Platform | 2025 Trading Volume | Core Advantages | Market Position | | Kalshi | ~ $23.8 billion | Compliance Path + Institutional Funds | Event Contract Exchange | | Polymarket | ~ $22 billion | Global Permissionless + Broad Coverage | Crypto-Native Liquidity Hub | | Second-Tier Total | ~ $1.25 billion | Vertical Exploration | Marginal Ecosystem |

Logical Conclusion: The core of prediction markets is not technology but the composite moat of liquidity and event design capability. Low liquidity platforms find it difficult to win through decentralized competition.

## IV. Why Will Most Prediction Markets Fail?

Historically, failed platforms did not lose due to technology but due to the microstructure of the market.

### 4.1 Treating Prediction Markets as "Event Casinos"

This mistake leads to: high-frequency noise overwhelming information traders, market-making funds unable to stay long-term, and unsustainable Sharpe Ratios. Successful prediction markets must provide structural advantages for information traders.

### 4.2 Mismatched Sources of Liquidity

Prediction markets do not need retail investors but rather: macro traders, policy researchers, industry experts, and risk hedgers. They provide information-driven trading flows rather than gambling flows.

### 4.3 Incorrect Settlement Frequency Design

If the market settlement cycle is too short, it will degrade into gambling; if too long, it will lose capital efficiency. The optimal range is usually for events with an information half-life of 2 weeks to 6 months, which corresponds to the real-world time window where "disagreements can form but are still verifiable."

## V. Vertical Track Analysis: Four High-Growth Sub-Directions

As the window for general prediction markets gradually closes, track opportunities are concentrating in vertical directions. Sports, creator economy, AI predictions, and social bot interactions have become the four fastest-growing sub-tracks.

### 5.1 Sports Track

Key Logic

Sports events inherently have high-frequency schedules and clear outcomes, making them easy to quantify and predict while forming a highly engaged user base. Platforms can quickly build trading markets and odds systems through middleware (such as Azuro Protocol), lowering technical barriers.

Representative Projects

Football.fun: Short-term TVL exceeds $10 million, with high user engagement.
Overtime: Combines community interaction with derivative trading to form a closed-loop ecosystem.
SX Network, Azuro Protocol: Provide public chain and middleware support for sports predictions.

User Behavior Characteristics

High-frequency participation, instant betting, and active trading around events.
User operations are easily influenced by community and social recommendations.
Preference for leveraged tools and short-cycle contracts, seeking rapid feedback.

Trends and Opportunities

In the next 1-3 years, the sports track will further professionalize: high-frequency derivatives, leveraged trading, and cross-chain aggregation will become standard configurations, forming a composite growth model of "sports predictions + community economy" through community and event ecosystems.

### 5.2 Creator Economy Track

Key Logic

Combining prediction markets with the creator economy directly empowers KOLs with market generation and revenue distribution. Users become community content producers while participating in predictions, forming a closed-loop ecosystem through creator revenue sharing, leading to significant viral growth effects.

Representative Projects

Melee: Offers a 20% creator revenue share, incentivizing KOLs to drive market generation.
Index.fun: 30% creator revenue, packaging prediction results into a "creator index," enhancing secondary trading and community participation.

Trends and Opportunities

In the future, the creator track will move towards indexation and platformization: platforms can transform creator influence into tradable assets through prediction indices, NFT incentives, and revenue-sharing mechanisms.

### 5.3 AI Prediction Track

Key Logic

AI is upgrading from an auxiliary tool to a core product, taking on market generation, event parsing, content production, and settlement functions. Through intelligent agents and Copilot, platforms achieve zero-cost creation, infinite supply, and automated settlement, significantly reducing operational costs.

Representative Projects

OpinionLabs: AI agents generate event markets and automatically settle prediction results.
BuzzingApp: AI-driven zero-human operation supports rapid event iteration and settlement.

Trends and Opportunities

In the next 1-3 years, AI will become a standard feature of prediction markets: automated market generation, intelligent settlement, event parsing, and risk control will all be AI-driven, giving rise to new high-frequency and highly intelligent products while attracting professional quantitative traders.

### 5.4 Social Bot Interaction Track

Key Logic

Lightweight front-end and social embedding lower user operation barriers, directly embedding prediction trading into Telegram, X platform tweets, or content wallets, forming a closed loop of "social equals trading."

Representative Projects

Flipr, Noise: Telegram Bots simplify complex trading operations with one-click ordering.
XO Market: Aggregates orders from multiple platforms, providing leverage and stop-loss features to meet professional user needs.

Trends and Opportunities

In the future, the social bot track will deeply integrate platform aggregators and leverage tools, achieving cross-chain liquidity integration and further expanding user coverage through social embedding, becoming the "growth engine" of prediction markets.

Conclusion: The rise of vertical tracks reflects the trend of prediction markets evolving from general information tools to "derivatization + data service + AI embedding + creator/social ecosystem." Each track forms a complete logical chain: market drive → user behavior → technical support → investment opportunity.

## VI. Breakthroughs for Small Prediction Markets

Even with extremely high industry concentration, small platforms still have several "blue ocean" opportunities to tap into:

### 6.1 Vertical/Niche Markets

Professional sports events, esports, industry KPIs.
Internal corporate prediction markets, professional association events.
Specific industry or regional policy events.

Logic: Deep or specialized events that mainstream platforms cannot cover can form high-value but low-trading-volume markets.

### 6.2 Data Productization + B2B Embedding

Not directly engaging in trading but turning price signals into API/index products to sell to funds or enterprises.
Core advantages include low regulatory risk + sustainable business models.

### 6.3 Experience Differentiation / Information Value Addition

Providing pre-analysis tools for predictions and community consensus mechanisms.
Making predictions "cognitive value-added rather than pure trading," increasing user stickiness.

Core Logic: Small platforms should avoid direct competition in liquidity and focus on high-value, low-scale scenarios + data output-based business models.

## VII. Investment Perspective: Structural Infrastructure is the True Betting Direction

Potential high-value directions include:

Prediction market data APIs (selling to quant funds).
Enterprise-level decision market SaaS.
Market making and risk mediation.
Probability index products (similar to VIX's Future Expectation Index).

The true moat will belong to those who control probability distribution rather than those who facilitate trades.

### 7.1 VC Actual Investment Direction Overview

|-----------------|--------------------------------|---------------| | Investment Direction | Representative Projects | Investment Motivation | | Compliant Exchanges | Kalshi | To create "event futures CME" | | On-Chain Markets | Polymarket, Augur | Information asset trading | | Infrastructure / Clearing / Tool Layer | The Clearing Co., TradeFox | Building market plumbing | | Social / Vertical Prediction | Manifold, FUN Predict, Azuro | Exploring new application forms |

### 7.2 Key Financing Signal Interpretation

The Clearing Company completed approximately $15 million in financing, with investors including Union Square Ventures, Coinbase Ventures, Haun Ventures, and Variant. This is a very critical signal: capital is beginning to treat prediction markets as a formal asset class that requires a clearinghouse.

Kalshi's valuation has risen to $5 billion; FanDuel and CME are also preparing to launch prediction market products to compete; by 2025, institutional funds are expected to account for about 55% of prediction market capital. This indicates that it is undergoing an evolutionary path similar to 2017 DEX → 2021 DeFi → 2024 prediction market tech stack.

## VIII. Future Trends and Evolution Directions

### 8.1 Mechanism Evolution: Deepening Derivatization

Prediction markets will gradually move from "event outcome prediction" towards high-frequency trading, structured options, and leveraged contracts. Typical paths include:

Short-cycle event contracts (e.g., Limitless 30-minute contracts) → high-frequency volatility trading tools.
Leveraged trading (Flipr 5x) → integration with DeFi leverage protocols, forming an on-chain derivatives ecosystem.
Interval predictions, spread arbitrage → gradually evolving into structured options and financial derivatives.

At the same time, cross-chain and cross-platform liquidity integration will become core competitiveness. Aggregators will merge order books from different platforms, providing optimal pricing and settlement solutions, similar to "prediction market 1inch."

### 8.2 Product Evolution: Data Service + AI Embedding

Prediction market prices have already reflected "event probabilities" and will become the core data source for institutional quant, asset allocation, and risk management in the future. Product forms will include:

Data subscriptions: Providing real-time market probabilities, top account behaviors, and arbitrage opportunities.
Indexing: Combining different prediction results into a "creator index" or "event index" for easier secondary trading or embedding in DeFi.
Visualization terminals: "Prediction market Bloomberg terminals" in the style of Polysights, directly providing strategy signals.

At the same time, AI will participate in market generation, automated settlement, content parsing, and risk control: automatically generating event markets (zero human intervention), intelligent settlement and odds adjustment, with AI Agents/Copilots participating in trading predictions.

### 8.3 Infrastructure Evolution: Modular and Composable

Prediction markets will resemble DeFi Legos more: market generation, settlement, liquidity, oracles, AI Agents, and other modules will be modular, allowing projects to be plug-and-play, lowering technical barriers, and supporting multi-chain deployment.

Gnosis CTF → Standardized asset issuance.
Azuro Protocol → Gambling middleware.
Polymarket / Kalshi → Core settlement layer.

Multi-chain deployment and cross-chain order aggregation will become standard: Base, Polygon, Solana, and other chains will serve as the foundational infrastructure for vertical tracks.

### 8.4 User Experience Evolution

Front-end interactions will evolve towards socialization, lightweight, and immediacy: Bots (Telegram/social platforms), one-click ordering, and leveraged operations embedded in content ecosystems. AI + intelligent oracles will reduce manual operations and costs, while automated settlement and intelligent event parsing will enhance platform scalability.

In the next 1-3 years, prediction markets will exhibit an accelerated development trend driven by "derivatization + data service + AI embedding + composable infrastructure." Evolving from mere information aggregation tools to a comprehensive entity of financial derivatives markets + data services + AI ecosystems + creator/vertical track integration. Investment value will concentrate on infrastructure modules, data services, vertical track applications, AI, and interactive layer innovations.