Prediction markets are enjoying their golden age. Polymarket's coverage of the presidential election made headlines, Kalshi's regulatory victory opened up new avenues, and suddenly everyone wants to talk about this "machine of world truth." But behind this frenzy lies a more intriguing question: if prediction markets are truly so adept at predicting the future, why haven't they become more widespread?
The answer isn't sexy. The problem lies in the infrastructure—in the US, in regulation (for example, Kalshi received approval from the Commodity Futures Trading Commission (CFTC), and Polymarket achieved offshore operations), but infrastructure issues remain widespread. Even in regions where prediction markets are legal, the same fundamental challenges persist.
The dominant platforms of 2024 addressed these issues by throwing money at the problem. According to Neel Daftary, a researcher at Delphi Digital, Polymarket invested approximately $10 million in market maker incentives, at one point paying over $50,000 daily to maintain liquidity in its order book. Today, these incentives have collapsed to a mere $0.025 per $100 trade. Kalshi spent over $9 million on similar projects. These are not sustainable solutions—they are merely band-aids for structural wounds.
Interestingly, the challenges hindering the development of prediction markets are not mysterious. They are well-defined, interconnected, and—for the right entrepreneurs—quite easy to solve. After speaking with teams in the field and analyzing the current state of affairs, we identified five recurring issues. Consider them as a framework, a common set of terms, to help us understand why prediction markets, despite their theoretically promising future, remain in the testing phase.
These are not just problems, they are roadmaps.
Question 1: The Liquidity Paradox
The fundamental problem lies in liquidity. Or more precisely, it's the chicken-and-egg dilemma that has led to most prediction markets becoming ghost towns.
The mechanism is simple. New markets launch with low liquidity. Traders face poor execution—high slippage and price shocks make trades unprofitable. They leave. Low trading volume scares away professional liquidity providers, as they need stable fees to cover the risk. Without liquidity providers, liquidity remains scarce. This cycle repeats itself.
Data confirms this. On Polymarket and Kalshi platforms, most markets have trading volumes below $10,000. Even the larger markets lack sufficient depth to attract meaningful participation from institutional investors. Any large position will result in double-digit price fluctuations.
The root cause is structural. In a typical cryptocurrency liquidity pool (like ETH/USDC), you deposit two assets and earn fees as traders make trades—the value of both assets is preserved even if the price moves against you. Prediction markets are different: you hold contracts that become worthless if the market fails. There's no rebalancing mechanism, no residual value—only two outcomes: half the assets go to zero.
Worse still, you'll get "harvested." As the market nears settlement and the outcome becomes clearer, informed traders know more than you do. They buy the winning side from you at favorable prices, while you're still pricing based on outdated probabilities. This "toxic order flow" continues to bleed market makers dry.
Polymarket switched from an Automated Market Maker (AMM) model to a central limit order book in 2024 for this very reason: the order book allows market makers to immediately cancel their quotes when they realize they are about to be trapped. But this doesn't solve the fundamental problem—it merely provides market makers with some defensive tools to mitigate losses.
These platforms circumvent this problem by paying fees directly to market makers. But subsidies cannot be scaled. This model works for flagship markets—presidential elections, major sporting events, and popular cryptocurrencies. Polymarket's election market has ample liquidity. Kalshi's NFL market competes with traditional sports betting companies. The real challenge lies in all other aspects: markets where prediction markets could function are hampered by insufficient trading volume to support millions of dollars in subsidies.
The current economic model is unsustainable. Market makers don't profit from price spreads; instead, they are compensated by the platform. Even protected liquidity providers with loss limits (maximum loss of 4-5% per market) require ecosystem subsidies to break even. The question is: how can providing liquidity be profitable without burning through cash?
Kalshi's success model is gradually becoming apparent. In April 2024, they brought in Susquehanna International Group, a major Wall Street market maker, making it their first institutional supplier. The result: a 30-fold increase in liquidity, contract depth of 100,000 contracts, and spreads below 3 cents. However, this requires resources that retail market makers cannot provide: a dedicated trading platform, customized infrastructure, and institutional-grade capital investment. The key to this breakthrough lies not in higher rebates, but in getting the first truly market-predicting institutional investor to recognize it as a legitimate asset class. Once institutions participate, others will follow suit: lower risk, benchmark pricing, and naturally, increased trading volume.
However, there's a problem: institutional market makers need to meet specific conditions. For Kalshi, this means obtaining approval from the U.S. Commodity Futures Trading Commission (CFTC) and clear regulatory requirements. But for crypto-native and decentralized platforms—the numerous platform developers lacking regulatory moats or large-scale operations—this path is not viable. These platforms face different challenges: how to launch liquidity without providing regulatory legitimacy or trading volume guarantees? For platforms other than Kalshi and Polymarket, the infrastructure issue remains unresolved.
What are entrepreneurs trying?
Quality-weighted order book rebates reward liquidity to improve trading—for example, by shortening trade times, increasing quote sizes, and narrowing spreads. While pragmatic, this approach doesn't solve the fundamental problem: these rebates still require funding. Protocol tokens offer an alternative—subsidizing liquidity providers (LPs) through token issuance instead of using venture capital, similar to the launch models of Uniswap and Compound. Whether prediction market tokens can accumulate sufficient value to sustain issuance in the long term remains unclear.
Tiered cross-market incentives provide diversified liquidity across multiple markets, mitigate risk, and enable more sustainable participation.
Just-in-Time (JIT) liquidity provision only provides funds when users need them. Bots monitor large trades in a pool, inject centralized liquidity, collect fees, and allow immediate withdrawal. This approach is highly efficient in terms of capital, but it requires complex infrastructure and fails to address the fundamental problem: the risk remains borne by others. JIT strategies generated over $750 billion in trading volume on Uniswap V3, but trading activity was primarily driven by well-funded participants, resulting in minimal returns.
Continuous combination markets inherently challenge binary structures. Traders are no longer confined to discrete "yes/no" options, but rather express opinions within a continuous range. This aggregates liquidity that was previously scattered across related markets (Will Bitcoin reach $60,000? $65,000? $70,000?). Projects like functionSPACE are building this infrastructure, although it has not yet been tested at scale.
The most radical experiments completely abandon order books. Melee Markets applies the Bonding curve to prediction markets—each outcome has its own curve, early participants get better prices, and those with strong conviction are rewarded. No professional market makers are needed. XO Market requires creators to inject liquidity using an LS-LMSR AMM, and market depth increases as funds flow in. Creators receive fees, thus linking incentives to market quality.
Both solve the cold start problem without requiring professional market makers. Melee's drawback is its lack of flexibility (positions are locked until settlement). XO Market allows continuous trading, but requires the creator to invest funds upfront.
Question 2: Market discovery and user experience
Even if the liquidity problem is solved, there is a more practical problem: most people cannot find the markets they care about, and even if they do, the experience is clunky.
This isn't just a "user experience issue," it's a structural problem. Market discovery problems directly exacerbate liquidity issues. Polymarket has thousands of markets online at any given time, but trading volume is concentrated in a very few areas: election markets, major sporting events, and trending cryptocurrency issues. Other markets are ignored. Even if a particular segment has some depth, if users can't find it naturally, trading volume remains low, eventually leading to market makers leaving. Vicious cycle: lack of market discovery means no trading volume, and therefore no sustainable liquidity.
Market liquidity is extremely concentrated. During the 2024 election cycle, Polymarket's top markets accounted for the vast majority of trading activity. Even after the election, the platform still sees $650-800 million in monthly trading volume, but this is concentrated in sports, cryptocurrency, and viral markets. Thousands of other markets—such as local issues, niche communities, and oddities—remain virtually untouched.
User experience barriers exacerbate this situation. The interfaces of Polymarket and Kalshi are designed for those already familiar with prediction markets. The average user faces a steep learning curve: unfamiliar terminology, the conversion between odds and probabilities, the meaning of "buy a YES," and so on. This is tolerable for cryptocurrency natives, but for others, these frictions stifle conversion rates.
Better algorithms can help, but the core issue is distribution: matching thousands of marketplaces to the right users at the right time without causing choice paralysis.
What are entrepreneurs trying?
The most promising approach is to offer services directly on users' existing platforms, rather than having them learn a new one. Flipr allows users to trade markets like Polymarket or Kalshi by tagging bots directly in their Twitter feeds. For example, when a user sees a market mentioned in a tweet, they can simply tag @Flipr to trade without leaving the app. It embeds market prediction into the conversational level of the internet, transforming social feeds into a trading interface. Flipr also offers leverage up to 10x and is developing features like copy trading and AI analytics—essentially, it's striving to become a fully-featured trading terminal, and that terminal happens to exist right on Twitter.
A deeper insight is that for startups, distribution is more important than infrastructure. Instead of spending millions of dollars like Polymarket to launch liquidity, it's better to consolidate existing liquidity and compete on distribution. Platforms like TradeFox, Stand, and Verso Trading are building unified interfaces that aggregate odds from multiple platforms, route orders to the best trading venue, and integrate real-time news feeds. If you're a serious trader, why bother switching between multiple platforms when you can use a single, more efficient interface?
The most experimental approach treats market discovery as a social problem, not an algorithmic one. Fireplace, part of Polymarket, emphasizes investing with friends—recreating the energy of collective betting rather than going it alone. AllianceDAO's Poll.fun goes a step further: it builds a P2P marketplace between small social circles where users can create marketplaces on any topic, bet directly with their peers, and have the creators or groups vote to determine the outcome. This model is highly localized and highly social, and by focusing on community rather than scale, it completely avoids the long tail problem.
These are not just improvements in user experience, but also distribution strategies. The platform that ultimately wins will not necessarily have the best liquidity or the largest market, but rather the one that best answers the question, "How can we deliver the predicted market to the right users at the right time?"
Question 3: Issues with users' expression of opinions
The following data should cause concern among all those who are bullish on the prediction market: 85% of Polymarket traders have negative account balances.
To some extent, this is unavoidable—prediction is inherently difficult. But part of the problem lies in the inherent flaws of the platforms. Because traders cannot effectively articulate their opinions, the platforms force them to build suboptimal positions. You have a meticulously crafted theory? No way. You're limited to binary bets: buy, don't buy, or choose position size. There's no leverage to amplify your convictions, no way to integrate multiple viewpoints into a single position, and no conditional outcomes. When traders cannot effectively express their convictions, they either tie up too much capital or under-leverage. Either way, the platform captures less traffic.
This problem can be divided into two distinct needs: traders who want to leverage a single bet, and traders who want to combine multiple perspectives into a single bet.
Leverage: Continuous Settlement Solution
Traditional leverage strategies are not suitable for binary prediction markets. Even if your prediction is correct, market volatility can wipe you out before settlement. For example, a leveraged "Trump win" position could be liquidated within a week of poor poll results, even if Trump ultimately wins in November.
But there's an even better approach: continuously settled perpetual contracts based on real-time data streams. Seda is building true perpetual contract functionality based on Polymarket and Kalshi data, allowing positions to be settled continuously rather than waiting for discrete events. In September 2025, Seda enabled perpetual contracts (initially with 1x leverage) for the live odds of the Canelo vs. Crawford match on its testnet, demonstrating the model's viability in sports betting.
Short-term binary options are another increasingly popular trading method. Limitless, which surpassed $10 million in trading volume in September 2025, offers binary options on cryptocurrency price movements. This type of market provides implicit leverage through its payout structure while avoiding the risk of liquidation for traders during the contract's life. Unlike fixed-income options, binary options settle at fixed intervals, but their immediacy (within hours or days, rather than weeks) provides the rapid feedback that retail traders need.
The infrastructure is maturing rapidly. Polymarket partnered with Chainlink in September 2025 to launch a 15-minute cryptocurrency price market. Perp.city and Narrative are experimenting with continuous news feed trading based on poll averages and social sentiment—true perpetual contracts that will never arrive at a binary outcome.
Hyperliquid's HIP-4 "Event Perpetual Contracts" are a groundbreaking technology—they trade on constantly changing probabilities, not just the final outcome. For example, if Trump's chances of winning rise from 50% to 65% after the debate, you can profit without waiting for election day. This solves the biggest problem with leveraged trading in prediction markets: even if the prediction is ultimately correct, positions can be forcibly liquidated due to market volatility. Platforms like Limitless and Seda are also gaining increasing attention with similar models, demonstrating that what the market needs is continuous trading, not binary betting.
Combination betting: An unresolved issue
Combination betting is different. It expresses complex, multifaceted assumptions, such as: "Trump wins the election, Bitcoin price breaks $100,000, and the Federal Reserve cuts interest rates twice." Sports betting companies can do this easily because they act like a centralized institution, managing decentralized risk. Conflicting positions cancel each other out, so they only need to stake on the maximum net loss, not on every individual payout.
Prediction markets cannot do this. They act as escrow agents—each trade must be fully pledged once it's completed. This quickly drives up costs: even for small portfolio bets, market makers need to lock up orders of magnitude more capital than sports betting companies would need to take on the same risk.
The theoretical solution is a net margin system that only stakes the maximum net loss. However, this requires a sophisticated risk engine, real-time correlation modeling across unrelated events, and potentially centralized counterparties. Researcher Neel Daftary suggests starting with a limited portfolio of market betting by professional market makers, gradually scaling up. Kalshi adopted this approach—initially offering only portfolio betting on the same event, as the platform could more easily model correlations and manage risk within the context of a single event. This approach is insightful but also acknowledges that a true portfolio market, the "choose what you want" experience, may be difficult to achieve without centralized governance.
Most prediction market entrepreneurs believe that these novel prediction market features have limitations: for example, leverage limits for short-term markets, pre-audited event combinations, or simplified "leveraged trading" that platforms can hedge. While the issue of user opinion expression may be partially addressed (e.g., continuous settlement), other aspects (e.g., arbitrary combination markets) remain far from being feasible for decentralized platforms.
Question 4: Permissionless Marketplace Creation
Solving the problem of market representation is one thing, but the deeper structural question is: who has the right to create the market?
Everyone agrees that the prediction market needs diversity—events with significant regional characteristics, events that attract niche communities, and unique, one-off events that traditional platforms would never touch, etc. However, creating a permissionless market has always been a challenge.
The core issue is that trending topics have a limited lifespan. The most explosive trading opportunities often arise during breaking news and cultural events. For example, a market for questions like "Will the committee revoke Will Smith's Oscar for slapping Chris Rock?" can generate huge trading volumes within hours of the event. But by the time centralized platforms review and list the news, interest has already waned.
However, creating a completely permissionless platform presents three problems: semantic fragmentation (ten versions of the same problem divide liquidity into useless pools), liquidity cold start (zero initial liquidity makes the chicken-and-egg problem extreme), and quality control (the platform is filled with low-quality markets, or worse—assassination bets that pose legal risks).
Both Polymarket and Kalshi opted for a platform curated selection model. Their teams audit all marketplaces to ensure quality and clear solution standards. While this helps build trust, it sacrifices speed—the platform itself becomes a bottleneck.
What are entrepreneurs trying?
Melee employs a strategy similar to pump.fun for its initial launch phase. The market creator receives 100 shares, with early buyers receiving progressively smaller shares (3, 2, 1, etc.). If the market gains traction, early participants can reap substantial rewards—potentially 1000 times or more. This is a "market of markets," where traders build positions early to predict which markets will grow. The core principle is that only the highest quality markets—those created by top creators or products truly meeting market needs—can attract sufficient trading volume. Ultimately, the best markets will naturally emerge.
XO Market requires content creators to provide liquidity using the LS-LMSR AMM. Creators earn revenue by paying fees, thus linking incentives to market quality. Opinion marketplace platforms like Fact Machine and Opinions.fun allow influencers to monetize their cultural capital by creating viral marketplaces around subjective topics.
The theoretical ideal is a hybrid, community-driven model: users invest reputation and liquidity when creating marketplaces, which are then vetted by community administrators. This model enables rapid, permissionless creation while ensuring content quality. However, no mainstream platform has yet successfully implemented this model. The fundamental contradiction remains: permissionlessness fosters diversity, while administrators ensure quality. Disrupting this balance will unlock the localized, niche markets that the ecosystem needs.
Question 5: Oracles and Settlement
Even if you solve all these problems of liquidity, discovery, expression, and creation, there is still a fundamental problem: who decides what happens?
Centralized platforms, where decisions are made by a team, are highly efficient but risk single points of failure. Decentralized platforms require oracle systems to handle arbitrary problems without continuous human intervention. However, determining the outcomes of these problems remains the most challenging aspect.
As researcher Neel Daftary explained for Delphi Digital, the emerging solution is a multi-layered stack that routes problems to the appropriate mechanisms:
Automated data delivery is used for objective results. Polymarket integrated Chainlink in September 2025 to enable instant settlement of cryptocurrency price markets. It is fast and highly deterministic.
AI agents are used to solve complex problems. Chainlink tested its AI oracle in 1,660 Polymarket markets, achieving an accuracy rate of 89% (99.7% accuracy in sports events). Supra's Threshold AI oracle utilizes a multi-agent committee to verify facts and detect manipulation, ultimately providing signed results.
Optimistic oracles like UMA are suitable for ambiguous problems; they propose several outcomes, and the disputing parties invest funds to challenge these outcomes. Although based on game theory, they are effective for well-defined problems.
For high-risk disputes, a reputation-based jury is used, with voting rights linked to on-chain performance records, not just capital.
The infrastructure is maturing rapidly, but market settlement remains the most challenging issue. If the settlement scheme goes wrong, it undermines trust; if it works, it can scale to millions of markets.
Why these questions are important
Liquidity, market discovery, trader opinion expression, market creation, and settlement are five interconnected issues. Solving liquidity problems enhances market attractiveness, thereby improving market discovery mechanisms. Better market discovery mechanisms attract more users, enabling permissionless market creation. More markets mean a greater demand for powerful oracles. This is a system, and currently, that system has bottlenecks.
But opportunities also arise: existing projects are trapped in established patterns. The success of Polymarket and Kalshi was built on certain assumptions about how prediction markets work. They were optimizing under given constraints. The new generation of developers, however, has the advantage of completely disregarding these constraints.
Melee can experiment with different bonding curves because their goal isn't to become a polymarket. Flipr can embed leverage mechanisms into social news feeds because they don't require regulatory approval in the US. Seda can generate perpetual contracts based on continuous data streams because they are not limited by binary parsing.
This is where the real advantage of entrepreneurs in the prediction market lies. It's not about replicating existing models, but about directly tackling the fundamental problems. These five problems are basic requirements. Platforms that can solve these problems will not only gain market share but also unleash the full potential of prediction markets as a coordination mechanism.
2024 proved that prediction markets can be adopted at scale. 2026 will prove that they can work everywhere.
Click to learn about BlockBeats' job BlockBeats.
BlockBeats to the official BlockBeats community:
Telegram subscription group: https://t.me/theblockbeats
Telegram group: https://t.me/BlockBeats_App
Official Twitter account: https://twitter.com/BlockBeatsAsia
