Author: Noveleader Source: Castle Labs Limitless Translation: Shan Ouba, Golden Finance

I. Industry Overview

Over the past 18 months, monthly trading volume in prediction markets has surged from approximately $2 billion to over $30 billion, evolving from a niche track into a mature ecosystem. Today, prediction markets serve as both information trading platforms and differentiated hedging tools, covering diverse categories such as sports, politics, and macroeconomics. Users can bet and trade on various events, from podcast commentary content to Federal Reserve interest rate decisions, with corresponding trading markets established for each.

The current industry explosion relies entirely on a platform curation model: the platform designates a small number of entities to decide which trading targets go online. Industry leaders Polymarket and Kalshi both adopt this model, strictly controlling the categories of markets available and reaping the market dividends.

Although the platform curation model has achieved a commercial closed loop, it deviates from the core vision of early prediction market products — permissionless, allowing anyone to create markets on any topic. Early products like Augur, Omen, and Zeitgeist, which championed permissionless creation, all failed. The core pain points are concentrated in four major issues: liquidity depletion, ineffective outcome resolution mechanisms, lack of creator incentives, and compliance risks.

The industry once concluded that the permissionless market model cannot scale. However, this conclusion is now being re-evaluated. A new generation of products attempts to solve the legacy problems of the first generation by optimizing oracle infrastructure, building scalable liquidity models, and redesigning products. This article reviews the development trajectory of prediction markets, analyzes the root causes of the initial failure of the permissionless model, the logic behind the success of the platform curation model, and how current new products are re-implementing the permissionless attribute of prediction markets. It also uses Limitless as a case study to interpret its newly launched User-Generated Market (UGM) feature's adaptation plan.

Early Prediction Markets: The Permissionless Approach

The power of prediction markets first gained attention when Polymarket and Kalshi made more accurate predictions about the 2024 U.S. election. Since then, the function of prediction markets has expanded far beyond that.

Since the beginning of this year, the nominal trading volume of prediction markets has consistently exceeded $20 billion, breaking through $40 billion last month, with the largest contributions coming from the two giants of the industry: Polymarket and Kalshi.

Prediction markets (PMs) were one of the earliest experimental primitives in the crypto space. Given the ethos of decentralization and permissionless participation in cryptocurrency, they initially adopted a permissionless approach.

Augur was one of the first products launched in this category in July 2018.

Founded in 2014, its ICO was a first in the field, raising $5.5 million.

In the Augur system, market creation was completely permissionless.

This permissionless architecture also brought its own problems. Within weeks of launch, Augur's user interface displayed trading markets for political assassinations, plane crashes, and other legally sensitive events. Additionally, issues extended to technical limitations, such as high gas fees on Ethereum and settlement delays (resolution could take up to 90 days due to disputes).

An interesting aspect of the product was its native REP token, where holders could stake the token to report outcomes.

Augur is set to restart in 2026 under the support of the Lituus Foundation, serving as dispute resolution infrastructure that other project management organizations (PMs) can utilize. Currently, they are competing in the infrastructure space. In its new dispute resolution mechanism, participants commit their own funds to support the final outcome. With each dispute, the committed funds increase, making it increasingly costly to maintain an dishonest position.

Gnosis was another early participant in the field, developing the Conditional Token Framework in 2017, which allowed anyone to create markets and transform event outcomes into tradable tokens. Polymarket later adopted this framework. Due to similar issues faced by Augur, including high Ethereum gas fees, scalability problems, and a lack of user-friendly tools, Gnosis eventually shut down its product.

Omen launched in 2020, using the same framework developed by Gnosis, offering permissionless Automated Market Maker (AMM) market creation. However, one of the challenges it faced was liquidity fragmentation: anyone could create markets for any niche, leading to hundreds of nearly identical markets for the same event, most of which had extremely low liquidity.

Additionally, they faced oracle-related issues. Omen used Kleros for dispute resolution, an external, decentralized oracle infrastructure. Kleros's crowdsourced jurors are incentivized to align with the majority opinion, which may not necessarily reflect the facts or market outcomes. Furthermore, the dispute resolution process was very slow, and gas fees were high.

Up to this point, the problems were mainly a combination of technical issues and liquidity problems, but the platform's economic model and design also played a significant role.

In early 2018, Stox launched a licensed prediction market for sports, finance, and news, raising $33 million through an ICO. The core failure of Stox was the lack of an effective profit model and token economic model. Although the platform charged fees, these were insufficient to maintain market maker incentives. Additionally, holders of the native STX token were supposed to receive platform revenue through fee sharing, but this mechanism was never effectively implemented, thus failing to attract more users and liquidity. From a technical perspective, Stox's resolution oracle was centralized and relied on the company itself, raising concerns about centralization.

Another permissionless platform, Hedgehog Markets, built on Solana, allows users to create their own markets. They introduced "risk-free prediction markets" using principal-protected deposits. Users could stake $100 or $1000 in USDC and receive game tokens corresponding to the staked amount. These game tokens were used for actual predictions, while the yield from the USDC pool was distributed to winners. While this model sounded unique, it reduced the potential gains users could obtain from the market, as only the yield was the maximum amount they could earn. This created asymmetry for users genuinely interested in participating and willing to take risks, as the yield was their maximum gain.

All these products highlight problems in different areas, exposing issues in liquidity, event management, problem resolution, regulatory compliance, and more.

Discoverability Issues

With permissionless market creation, users can create any number of markets, which can lead to multiple identical markets, resulting in hundreds of markets for each major event with low liquidity, strange wording, and partial redundancy.

This redundancy creates a poor user experience, confusing most users. Experienced users may funnel their capital into the most liquid market, but the liquidity insufficiency can only be thoroughly resolved in two ways: either do not create similar markets, or build a front-end that effectively excludes low-liquidity markets, making them only accessible through search.

Dispute Resolution

"The effectiveness of a prediction market depends on its ability to discern the truth."

Both the Augur REP staking oracle and Kleros (Omen) incentivize stakers to vote in accordance with the majority. Similarly, in UMA, the optimistic oracle (used by Polymarket) faces the same resolution challenge as any market dispute: UMA token holders vote to make the final decision, and these voters themselves can also trade on Polymarket, leading to final decisions that favor their positions.

For example, on the Polymarket platform, the current dispute resolution mechanism is dominated by 9 "whale" addresses, whose rulings consistently align with the final winning outcome. These addresses can easily manipulate the market in favor of their own trades.

A recent example is the MicroStrategy BTC trading market. Given that Strategy sold 32 BTC in this market between May 26 and 31, the market should have been judged "Yes," but even after two appeals, the final result was "No."

__

Liquidity Fragmentation

The permissionless content aggregation model leads to liquidity dispersion, as users can aggregate content across multiple competing markets for the same topic, spreading liquidity thin. For Omen, the biggest challenge was liquidity fragmentation, as users aggregated content markets that were nearly identical.

This can only be solved by stricter curation processes, e.g., allowing only one market per event; or through the front-end, where on the interface, one topic and all different markets are displayed in the same location, allowing users to choose which market to trade while the system recommends the most liquid one.

Attack Surface

PMs are essentially susceptible to economic or informational manipulation.

Users can buy in large quantities to gain significant voting power on the oracle, thereby significantly altering market prices and steering decisions in their favor.

This problem exists in both licensed and permissionless models, but the attack surface expands much faster in the latter as users continuously create markets.

Recently, a Google employee was caught by the U.S. SEC for trading stocks related to Google searches on the Polymarket platform. He exploited information arbitrage, as he knew the correct stock price in advance, making millions of dollars in profit.

Regulatory Issues

The U.S. Commodity Futures Trading Commission (CFTC) treats event contracts as futures contracts, which require registration with a Designated Contract Market (DCM) under the Commodity Exchange Act.

Many older PMs did not comply with regulations, which was a major problem for scaling.

Due to non-compliance, Polymarket was fined $1.4 million in 2022. The U.S. market is huge, with active sports betting and trading. For this reason, Kalshi has always adhered to a regulatory-first strategy, and Polymarket also acquired QCEX in 2025 to obtain U.S. regulatory approval.

No Curator Economics

Early permissionless market management models did not pay much attention to the economic benefits of market management, nor did they consider how managers could profit from building markets. This led to an asymmetry between the platform and market managers, as even if market popularity and liquidity increased, managers lacked motivation and could not obtain economic benefits from it.

However, some of these issues were understood by emerging platforms (now industry giants), who solved them by adopting a licensed content management approach.

Licensed Curation: The Polymarket and Kalshi Approach

Polymarket launched in 2020, while Kalshi launched in 2021. These two platforms took different approaches, but recently they have begun to converge in terms of regulatory compliance and winning the U.S. market.

Over time, Kalshi and Polymarket have become strong competitors.

Until mid-2025, Polymarket maintained an 80% market share. After that, Kalshi established its dominance through partnerships with platforms like Robinhood, and now accounts for the majority of Polymarket's trading volume.

In terms of fees, Kalshi's current annualized fee revenue is approximately $2 billion, while Polymarket's is about $300 million. Additionally, in recent funding rounds, Kalshi and Polymarket were valued at $22 billion (Kalshi's current target valuation is $40 billion) and $15 billion, respectively, and by price-to-sales ratio, Polymarket appears overvalued relative to its competitor**.**

To reach this scale, Kalshi and Polymarket have solved most of the problems we mentioned with early prediction markets.

Starting with regulatory compliance, Kalshi's regulatory-first strategy has paid off handsomely and helped it grow in the U.S. Polymarket** is also following suit with Polymarket US**, which is being rolled out in phases.

In terms of market management, both platforms have active market management teams and are licensed, which helps prevent issues like liquidity fragmentation and duplicate markets.

While the licensed nature of these platforms has been a successful approach so far, this does not mean that the permissionless model cannot coexist in this category, as many users still need control over markets, creating niche markets on topics of interest to them, sharing curator revenue, and participating in building liquidity.

The permissionless model and the early versions of PMs exposed many problems, which incumbents have set out to solve.

Even though existing markets have been around for a while and have reached sufficient scale, users still cannot create and trade markets of their choice, and the ultimate goal of prediction markets is to become the source of truth for any topic backed by economic interests.

Given this, permissionless content management is experiencing a massive revival as the solution space expands.

In the following sections, we will discuss how permissionless market creation becomes a category and how User-Generated Markets (UGMs) expand.

The Permissionless Pivot: A New Generation of PMs

**Over the past year, the permissionless content curation space has grown and evolved, with new players **(Melee, HIP-4, XO Market, etc.) entering the market, and established products like Limitless also entering the space as they expand their business to offer User-Generated Markets (UGM).

Limitless recently launched permissionless trading markets. They have not yet fully opened permissionless trading management; **initially **only crypto-related trading markets are allowed, with gradual expansion to other categories later, in order to better understand market demand and scale accordingly.

Additionally, curators on the platform receive 50% of the revenue from the markets they curate, creating a direct interest alignment between the platform and market curators .

HIP-4 super-liquid bounded markets adopt a capital threshold mechanism, requiring market deployers to stake 1 million HYPE tokens to obtain a deployment slot. Builders can choose to charge an additional fee split of up to 50% on top of Hyperliquid's base fee (currently zero to encourage trading, as the market launched in early May). This mechanism helps Hyperliquid thoroughly eliminate junk markets, as the threshold for deploying outcome markets is very high.

Another structural advantage of HIP-4 markets is that they run natively on Hypercore, sharing the same order book, account structure, and margin engine as Hyperliquid's spot and perpetual markets, meaning traders can access prediction markets and hedge their portfolios without having to spread funds across different accounts.

Both of these platforms support order books, with liquidity provided by market makers (MMs).

Other platforms adopt different strategies for liquidity.

XO Market uses the Liquidity-Sensitive Logarithmic Market Scoring Rule (LS-LMSR) AMM model, an improvement over the standard LMSR used in most prediction markets.

The key difference lies in how liquidity is managed: in standard LMSR, market managers must pre-set a fixed liquidity parameter, essentially estimating how much trading activity the market will attract. If the **parameter is set too low, prices are too sensitive to trades; setting it too high requires committing too much capital. LS -LMSR eliminates this burden by dynamically adjusting the liquidity parameter, allowing market depth to automatically adjust as trading activity accumulates.

It also requires users to provide initial liquidity when creating a market, which helps discourage spam and reduce illiquid markets, so it is hailed as a "conviction market," since users must provide liquidity.

In terms of dispute resolution, XO uses a unique three-tier system. The first tier is an AI-first path, using the MODRA (Market Outcome and Dispute Resolution Agent) system to leverage AI for fast, autonomous resolution of clear-cut cases. This is followed by appeals to the Senate Court and the Supreme Court, which require human review. To date, the system has processed over $250 million in trading volume, supporting more than 2,800 markets and over 30k transactions.

Melee is another PM on the list, which is expanding its pari-mutuel betting market model, which they call "PMM". In traditional pari-mutuel betting markets, all bets are pooled together, and payouts depend on the total wagered amount and the number of winners. Although traders cannot enter or exit after a specific time period, this model is commonly used in horse racing; after the race starts, bettors cannot exit.

Pari-mutuel betting markets bring a series of problems, such as requiring continuous liquidity, because **users can only exit after the market **settlement . The inability to exit during the market leads to trader loss due to lack of flexibility. Additionally, to exit, traders may need to buy the opposite tokens while the market is open, and given the nature of pari-mutuel markets, the market value can change at any time. Moreover, under this design, the market needs to close before the event starts, because during the event, market participants might only buy tokens for the winning side, thus manipulating the market mechanism.

Melee is actively working to adopt the pari-mutuel betting model by enabling continuous trading to remove limitations, but has not yet officially launched.

Products like Xmarket handle initial market liquidity differently from others.

The minimum seed capital required for market creation is $1, but a market only officially goes live once the initial liquidity reaches a soft cap of $100. If any market fails to reach this cap, all participants receive a refund. This effectively filters out spam, screening for markets that attract genuine attention and judging whether users are interested in the market. However, because this threshold is low, it can be gamed: a market creator can push for the soft cap to be reached, allowing the market to go live, but this still helps kickstart initial liquidity and activity.

In the chart below, we **more specifically clarify the differences between these products by comparing the problems encountered by first-generation PMs **. Multiple solutions exist for key issues such as curation, dispute resolution, liquidity, economics, and regulatory paths, indicating a willingness to experiment in the space.

UGMs have faced many challenges in the past, and there is still no conclusive evidence on which approach works best, as no UGM has reached significant scale yet.

In the next section, we will detail one of the recently launched products, Limitless, focusing on its approach to User-Generated Markets (UGM) and the problems raised in the first section.

Financial Prediction Markets: How Limitless UGM Works

Limitless has been active in the PM space since 2024 and has grown steadily. This growth can also be attributed to their use of tokens as a distribution tool for user incentives.

The Limitless platform uses a licensed content management process, with all markets designed internally by the team. The settlement process is also carefully designed, with Pyth and Chainlink serving as oracles for financial markets like crypto, stocks, commodities/forex, while sports, politics, and other categories handled by the Limitless team use manual settlement. If a market cannot be settled, users receive a refund.

Given the infrastructure they have built, they are expanding their services into the permissionless content aggregation space.

On June 2, 2026, Limitless launched its first permissionless trading market type. To scale, they deliberately adopted a template-based financial market trading model, rather than opening trading for any topic. In the initial phase, creators can choose from a fixed set of crypto assets (BTC, ETH, SOL, XRP, and DOGE), set a bounded price target (-5% to +5%), and select a trading duration from 15 minutes to 1 day. The purpose of the templated approach is to avoid markets with ambiguous or obscure wording.

In the first month after launch, these markets achieved a trading volume of $2.2 million.

**User-Generated Markets (UGM) operate in parallel with the platform's internally curated markets, **making Limitless a hybrid platform. Additionally, only crypto trading is supported in the initial phase; the team will gradually add other trading categories based on market demand and platform standards.

As these markets scale, User-Generated Markets (UGM) can evolve into customized hedging tools, allowing users to build their own markets based on their positions (including spot and forward positions). As more asset classes and PM categories like sports and esports are added to the platform, their use cases will further expand.

Limitless adopts a unique approach to solving the unique problems faced by early PM implementations.

Creation Interface: Limitless starts with a limited scope, focusing on a few financial markets, starting with crypto (BTC, ETH, SOL, XRP, and DOGE). Users select price and market duration within a templated configuration, reducing the likelihood of settlement issues due to non-standard market descriptions.

Resolution: Since they initially focus on crypto-related markets, they use the same oracles as their licensed product. They leverage services like Chainlink and Pyth to maintain consistent price data on the platform. The presence of automated oracles eliminates the need for continuous monitoring, making market creation and closure smoother.

Liquidity: Limitless provides initial liquidity. CLOB creates UGMs to facilitate an open infrastructure, allowing any market maker to enter and provide liquidity, enabling efficient price discovery.

Curator Incentives: For the success of User-Generated Markets (UGM), the economic model for curators and how they profit throughout the process is crucial. Because without a substantial source of income, curators will not curate within normative boundaries, creating many uncertainties that are detrimental to market discovery and volume growth. This balance between the platform and curators is vital for maintaining aligned interests. On the Limitless platform, curators pay a fee of 100 to 1,000 LMTS (depending on market duration; longer duration, higher fee) to create a market and receive 50% of the fee revenue from their curated markets. Additionally, the team is exploring different pricing models based on user feedback.

Discoverability: Since markets are only for the prices of a few assets, the possible market space is limited in terms of assets, durations, and price targets, constraining market fragmentation and near-duplicate markets.

Regulatory Compliance: In early May, Limitless filed an application with the U.S. Commodity Futures Trading Commission (CFTC) for approval to operate as a federally regulated derivatives exchange in the U.S. (the CFTC has deemed the application substantially complete and has entered formal review). If approved, this would provide a clear regulatory path for Limitless to expand its prediction market business in the U.S., actively competing with Kalshi, Polymarket US, and crypto.com derivatives.

Limitless's approach offers an effective solution set for the various problems that emerged from early PM versions. Its templated configuration maintains platform integrity and alleviates issues related to ambiguous markets. While Limitless charges a fee for creating markets, this effectively addresses the spam problem, as setting a threshold for market creation reduces the cost for spammers. The 50% fee sharing truly incentivizes creators to participate and attract active users. Furthermore, Limitless is awaiting CFTC approval, and its clear regulatory path sets it apart from peers.

However, the platform's fee also creates a "chicken and egg" problem: it requires users to pay a fee to create these markets, so users must weigh their investment against the potential returns from the 50% market fee share. The latter only increases if trading activity is sufficiently high, while the former remains fixed.

In the current Limitless model, users pay a fixed upfront fee, denominated in its native token LMTS. While this helps prevent spam, the platform could consider adding dynamic fees based on content demand, so that fees can be lowered when demand for a particular market or asset is low, and vice versa.

What Happens Next?

PMs appeared early, with early products like Augur and Gnosis launching during the crypto ICO era. These products faced very similar problems, primarily centered on liquidity fragmentation, slow or disputed settlements, weak economic incentives for PMs, and regulatory risks.

The new wave of products in the permissionless content curation space recognized these gaps and adopted unique designs to address the problems currently faced by PMs.

XO Market allows creators to bootstrap initial liquidity and adds a tiered settlement model for the most accurate outcomes. Melee redesigned the pari-mutuel betting model to enable continuous trading. Xmarket attempts to filter spam through a soft capital cap. HIP-4 sets a high staking threshold for creation to ensure only high-quality markets exist. Limitless introduces a 50% fee share and curation fees to improve curator economics and prevent spam.

But what truly differentiates them is not the solutions they offer, but their completeness across different dimensions. A product that solves the liquidity problem without addressing the capital flow problem, or solves the capital flow problem without improving creator economics, will not be able to scale.

Given this, permissionless markets are being re-examined, and it is worth paying attention to.

Over time, it will become apparent which design choices scale best and become the best place to develop markets in any specific niche, and to find the truth, PMs should look to architecture for the answer.