Author: Will Owens, Galaxy Digital research analyst; Translation: @金财经xz

The prediction market has entered a new stage of mainstream vision and capital formation. What was once a niche area of ​​on-chain speculation is quickly becoming one of the fastest growing segments in the financial industry.

The significant growth of Polymarket (daily trading volume hit a new high, valuation reached US$9 billion, and was recognized by traditional finance) and the success of Kalshi (top of the list of free financial apps in the iOS App Store) jointly prove that the product-market fit that the prediction market has long pursued has been achieved.

Prediction markets date back to at least the 16th century, when true Renaissance elites were placing bets on papal elections. The prototype of modern prediction markets originated from the idea of ​​Robin Hanson, an economist at George Mason University. Hansen's early work on information markets and "voting markets" illustrated how financial incentives could be used to aggregate dispersed information more effectively than polls, forecasts, or expert committees. His ideas laid the conceptual foundation for prediction markets—considering them as probabilistic price discovery tools rather than mere gambling products. This is also the current market positioning strategy of Kalshi and Polymarket.

This is not gambling, this is prediction. Is it really so?

Currently, the competitive landscape of the prediction market is broadening. A new round of entrants are accelerating experiments in product design and liquidity incentive mechanisms. Decentralized finance protocols are also exploring the potential integration of predictive assets. In theory, binary outcome shares (Yes/No tokens) can function as composable financial instruments outside of their native prediction markets. For example, "Yes" shares tied to the election results can be used as collateral for long Bitcoin positions in the perpetual contract DEX.

At the same time, clarity on the regulation of event contracts in the United States has opened the door for the parallel expansion of regulated platforms and fully on-chain venues. Please refer to Polymarket’s newly launched US-compliant mobile application (currently only supports sports betting).

Since the 2024 U.S. presidential election, prediction markets seem to have become the core of all discussions about information discovery and on-chain finance. More and more companies are launching their own prediction market products, and the influx of capital is rising sharply. The field is undergoing a major transformation. This report will explain the current state of the development of prediction markets, the core innovations driving its evolution, and why the next stage may blur the line between event markets and derivatives.

Core points:

The prediction market has crossed the chasm and entered the stage of mainstream awareness and capital formation: Polymarket and Kalshi rank among the fastest-growing consumer financial products; Polymarket’s cumulative number of unique users has exceeded 1.6 million; the Kalshi mobile application ranks first among iOS financial applications.

The clarification of U.S. regulations is promoting the simultaneous expansion of prediction market models domestically and overseas exploration.

Prediction markets are evolving from mere betting tools to fundamental financial components.

Leveraged prediction markets may develop into event risk hedging products similar to perpetual contracts.

Artificial intelligence will reshape the pricing mechanism, usage scenarios and cognitive methods of the prediction market.

Liquidity remains a core constraint in the sector.

Brief description:

At present, many readers may have understood the operating mechanism of the prediction market and its core value - extracting effective signals from complex information. If you are still unfamiliar, here is a brief explanation.

Take "Will the Buffalo Bills win the Super Bowl?" as an example: users can purchase "Yes" or "No" shares. If the Bills win, each "yes" share will be paid $1, otherwise it will be $0; the settlement method for "no" shares is reversed.

The share price reflects the probability of the event occurring. If the "yes" share trades at $0.22 and the "no" share trades at $0.78, it means the market thinks the Bills have a 22% chance of winning the Super Bowl.

Theoretically, the greater the trading volume and liquidity these markets attract, the more accurately their prices can gather the wisdom of the crowd to form a reliable estimate of the true probability.

1. Prediction markets serve as hedging tools

One of the most underrated use cases for prediction markets right now is hedging crypto tokens before they go live. Take the market where the fully diluted valuation on the first day of token listing is used as an example for settlement. "What will be the fully diluted valuation of Monad on the first day of listing?" on Polymarket is a typical case. For traders and participants who hold MON quotas or pre-sale positions, such instruments often provide a purer hedging solution than the pre-sale “super perpetual contracts” traded on Hyperliquid that are designed to gain leveraged directional exposure to MON. This is mainly because they are able to effectively avoid the risk of short squeezes.

As we have pointed out previously, while pre-sale perpetual contracts allow traders to establish leveraged directional exposure before a token is issued, it also introduces liquidation risk. Even if a trader's long-term judgment or hedging direction is correct, a coordinated short squeeze may trigger forced liquidation before the hedging event occurs.

The Polymarket pre-sale market completely avoids this failure mode. Since there is no leverage or liquidation mechanism, the only possibility of losing principal is if preset conditions are not met at settlement. The trade either gets paid out or it doesn't. From a risk management perspective, this makes the Polymarket pre-sale market closer to binary options with discrete outcomes than to continuous derivatives.

Take the recipient of the Lighter airdrop as an example: if he expects to receive tokens when the mainnet goes online, and wants to hedge the downside risk on the first day of trading. The Polymarket market based on FDV settlement on the first day of LIT token issuance can provide risk exposure that is not disrupted by temporary price fluctuations or short squeezes. This type of hedging is precisely aligned with the core of the risk that participants are focused on: specific valuation checkpoints at clear time points with clear and verifiable settlement criteria.

This does not mean that the Polymarket pre-sale market is risk-free. Participants must have a thorough understanding of settlement standards and order book liquidity needs to meet position sizing requirements. Typically, though, these crypto pre-sale markets are liquid enough for most traders.

2, Prediction market current pattern

Prediction markets have grown from a niche on-chain novelty to the fastest-growing vertical in cryptocurrency. Although the current market size is still far from its theoretical potential, the foundation to support a larger-scale system is being built.

Economist Hansen described in a recent article that prediction markets are still in a "very early stage" compared to the world he envisioned in the 1990s. He warned that regulatory or cultural backlash against betting practices, insider advantages or celebrity-driven markets could stall or delay that trajectory for years. Galaxy Research predicts there will be federal investigations into prediction market insider trading in 2026.

However, this is not a black and white issue. Although the insider "problem" is often criticized, proponents (including Hansen) argue that insider trading is actually a "function rather than a bug" of prediction markets because it creates financial incentives for those who hold important non-public information to disclose it to the public. In fact, companies have tried to set up internal prediction markets where employees place bets on when product launches will happen, in an effort to reveal the truth rather than cater to management’s preferred information.

The current landscape is mainly defined by two dominant players. Although there are other players (including Opinion, which is backed by Binance founder Changpeng Zhao’s YZi Labs), Kalshi and Polymarket are still the undisputed leaders.

Polymarket:

Polymarket is the strongest proof that event contracts can achieve true consumer-grade scale. It is both a trading platform and a real-time probabilistic information flow. Its price is now regarded as a "definitive number", much like the traditional reference for odds, poll averages or implied interest rates. "Go and see Polymarket" has become a common phrase when discussing whether an event will happen.

Two forces have driven the platform’s growth since the 2024 U.S. presidential election cycle. First, the "signal market" (political, macroeconomic and high-profile real-life events) consolidates Polymarket's role as a probability barometer. Second, the expansion of non-political categories such as crypto-native events and sports not only increases coverage of tradable assets, but also improves user retention by providing tradable varieties on a daily basis, rather than just major news cycles. This category expansion is critical because prediction markets are ultimately subject to cadence constraints: the more frequently a market is listed, settled, and circulated with attention, the more "liquidity minutes" the platform can accumulate.

As a result, Polymarket has become the global leader in crypto-native prediction markets. Its daily nominal trading volume often reaches tens of millions of dollars, especially during political and macroeconomic events. The platform has become more than just a place for trading, it has evolved into a real-time feed of information for observers. When Intercontinental Exchange (ICE) announced a strategic investment of up to US$2 billion, it stated that it planned to explore the commercial development of Polymarket data (this investment valued the platform at approximately US$9 billion, making founder Shane Copeland the youngest self-made billionaire in the world), further solidifying this role.

User adoption will accelerate significantly after the 2024 U.S. election cycle, driven by viral market hot spots and the expansion of non-political categories such as sports and crypto-native events.

In December, the platform launched its first real-money trading mobile application for U.S. users after receiving approval from the U.S. Commodity Futures Trading Commission to become a federally regulated intermediate event contract trading venue. The app is currently only open to users in the sports market in the United States. It is gradually open for download through the iOS waiting list, and the Android version will be launched later.

Although its US business is currently limited to the sports field, Polymarket has stated that as the regulatory environment develops, it plans to expand into proposition and possible election markets. This marks Polymarket’s substantial return to the U.S. market after reaching a settlement with the CFTC in 2022 that prohibited it from doing business with Americans, not to mention the brief investigation storm it encountered at the end of the Biden administration.

In the entire ecosystem, Polymarket and Kalshi remain the core sources of mainstream capital flows. Kalshi dominates activity in US regulated markets, while Opinion is starting to generate significant weekly trading volumes. The chart below shows the notional amount of weekly spot transactions on each platform.

Kalshi：

Crypto natives often sneer at Kalshi because of its attributes of "operating off-chain" or being a "non-encryption company". In fact, the company's stance is more pragmatic than philosophy-driven. The team initially hoped to build an on-chain platform to gain the advantages of transparent settlement and composability, but regulatory uncertainty in the United States a few years ago made it difficult to advance this path without endangering the core business. Kalshi ultimately chose a completely off-chain architecture, aiming to be the first to launch compliant and scalable products, and expected to gradually add "on-chain" integration at the appropriate time.

This long-term openness is critical, as a number of "bridge" integration cases between Kalshi and crypto-native products have emerged over the past few months, such as:

DFlowTokenization (more on this below)

PhantomWallet Integration

Jupiter Prediction Market Integration

Coinbase platform access

These workarounds successfully bring on-chain liquidity and developer innovation to Kalshi without attempting to migrate entire regulated exchanges to the blockchain. Kalshi only needs to allow builders to encapsulate or connect their contracts in a way that DeFi and crypto native users can understand, and it will be able to access all on-chain liquidity at the same time.

The most striking thing about our conversations with the Kalshi team is that they purely view liquidity as a core constraint. Kalshi is now a mainstream product—as evidenced by the fact that its app topped the app store charts. However, most markets still lack sufficient liquidity to accommodate hedge funds and other mature capital to calmly arrange transactions (except for a few markets, because Wall Street giant Susquehanna International Group serves as its market maker). In Kalshi's vision, the "next phase of growth" does not require a completely new product, but a small baseline improvement in core market liquidity. This focus extends to all aspects, including internal plans to improve execution quality as transaction volumes grow.

The liquidity-first strategy also explains why the sports sector occupies an important position in Kalshi's business. Users will always trade what they want to trade, and sports is often the largest market. In short, sports have the broadest coverage of events, with consistent frequency and clear settlement criteria (even if fans boo the referee, the score is not disputed). No other category has comparable target breadth and mindshare. Although Kalshi's grand vision is to encompass all event futures (its company name means "all things" in Arabic), sports remain a natural liquidity magnet.

If we completely eliminate sports data and observe "non-sports" weekly transaction volume, we can obtain an analysis perspective with more reference value. Excluding the sports sector, it can be seen that Polymarket is still a more mainstream trading venue in terms of political/macroeconomic capital flows. In other words, non-sports trading volume is the fairest apples-to-apples comparison dimension for measuring a "signal market" because it strips away the ecological layer of trying to evolve into a financial infrastructure rather than a mass entertainment product.

Customer acquisition distribution and user experience are the other two major levers that Kalshi emphasized. The team emphasized that direct retail traffic is an important moat: if it can continue to acquire and retain consumers, it can gradually accumulate liquidity through "scale crushing" - which is often difficult to achieve by offshore trading platforms and weak on-chain order books. Because of this, Kalshi pays close attention to the marketing strategies of sports betting industry giants such as FanDuel and DraftKings, and continues to focus on product iteration.

Currently undergoing a large-scale user experience upgrade, including designing differentiated interfaces for traders of different maturity levels (beginners, intermediate, advanced). This is similar to the "Easy" and "Advanced" modes of the Coinbase mobile app (or to a ski resort offering multiple levels of slides from kiddie trails to black diamond tracks). The underlying logic is simple: the easier a prediction market is to trade, the more practical it will be; and a more convenient trading experience will attract more liquidity, making the price more reference-worthy - this is essentially a self-reinforcing flywheel of growth.

Kalshi’s roadmap views the ecosystem as a modular structure. The team painted a future picture of different front-ends and tools serving different user groups: native applications, professional terminals, Telegram robots and customized interfaces will all promote the maturity of the prediction market trading ecosystem.

We have witnessed exactly the same development trajectory in the memecoin space: initially traders manually redeemed through the Raydium pool, then using BONKbot, then Trojan, then GMGN, and finally Axiom launched the "ultimate" memecoin trading terminal. At the current stage, Kalshi’s top priority is still to implement the core strategy: increasing the liquidity scale and transaction volume by an order of magnitude.

3, new generation ecology

Every key opinion leader on crypto Twitter seems to be wearing a Kalshi or Polymarket certification badge these days. These companies are indeed investing heavily in bringing in social media influencers to promote their products. The current market narrative is that crypto traders are rotating from meme coins to prediction markets. As we discussed on the Galaxy Grid Podcast, prediction markets may be more sustainable than memecoins (although users of the former cannot count on turning $10 into $100,000 in a single transaction).

In this context, a new generation of prediction market platform and tool ecology is taking shape. The following lists some representative projects (not a complete list):

Opinion: Prediction market powered by YZi Labs

Augur: A reboot of the early Ethereum prediction market

Lightcone: Influence Market Platform

Azuro

PrediX

Limitless: Focus on fast expiry markets

Myriad Markets

Noise.xyz: for trading attention, narrative and mindshare

Hilo

Gondor: Lending DeFi protocol for Polymarket positions

Space: Leveraged prediction market on Solana (up to 10 times)

Melee: A hybrid of Polymarket and Pump.fun, using bonded curve liquidity.

Fliprbot: User Interface and Trading Terminal

Polycule: Telegram trading robot

MetaDAO: Market-driven governance protocol

Cue.market: pop culture, fashion, entertainment fields

Bullpen: Added Polymarket terminal function

Polyoptions: Polymarket share options

Most new entrants are not simply copying the models of Polymarket or Kalshi, but are exploring new infrastructures, such as:

• Leveraging event results

• Forecast market share of lending

• Decision-making market

• Opinion Market

• Multi-result unified market

• Influence Market

DFlow: Tokenized Kalshi

Another different but equally important direction is to directly tokenize the off-chain prediction market into DeFi. As an infrastructure protocol, DFlow's prediction market API encapsulates Kalshi's regulated event contracts into SPL tokens (the token standard equivalent to Ethereum ERC-20 on Solana), allowing these positions to enter on-chain financial applications for trading and integration while retaining off-chain settlement and compliance. Each Kalshi position thus becomes composable with on-chain applications.

At the technical level, DFlow uses a parallel liquidity program to connect off-chain Kalshi liquidity with on-chain transaction intentions: users issue orders in Solana, off-chain liquidity providers execute transactions, and the protocol then mints or destroys tokens representing the corresponding predicted market risk exposure. When the market settles, settlement funds flow back through a parallel liquidity program and winning tokens are redeemable for stablecoins.

Gondor: Introducing lending function

The essence of Gondor is to build a credit layer on top of Polymarket. Since each Polymarket position is an ERC-1155 token, these positions can be considered collateralized assets and can be leveraged and risk managed like other DeFi positions.

As mentioned above, if the corresponding result is beneficial to the holder, these positions can be exchanged for USDC at a 1:1 ratio.

Gondor allows traders to use these ERC-1155 positions as collateral into a lending vault built on Morpho (a blue-chip lending protocol that manages billions of dollars in deposits and has a long-term audit trail).

After depositing funds, traders can borrow up to 50% of their value in USDC based on their positions, and the borrowed funds will flow directly back to the Polymarket platform and be displayed as the "cash balance" of the account. This allows traders to release idle equity locked in profitable or positive mark-to-market positions and redeploy capital to new trades while maintaining exposure.

Please note that not all markets are eligible. The most significant risk is that illiquid markets, of which there are many, may be subject to targeted liquidations and market manipulation. The Gondor team will manually review the Polymarket market supported by the lending agreement, and sometimes even only support unilateral positions (for example, only "yes" shares are supported) if the liquidity on the corresponding side is too scarce. Market whitelisting is reviewed based on several factors: order book depth, clarity of settlement criteria, and remaining settlement time. Each supported market is mapped to a specific lending vault, each with independent risk parameters, loan-to-value ratio caps and loanable exposure limits. Borrowing capacity is limited by the deposit size of Gondor's vault and the depth of liquidity in the underlying Polymarket order book. The borrowing limit is set based on the principle that positions can be reasonably hedged or closed in the underlying Polymarket market.

The maximum borrowing limit is set at a loan-to-value ratio of 50%, and the liquidation threshold is approximately 77% LTV. Positions established with the highest leverage will need to fall by approximately 35% in collateral value before liquidation is triggered.

For example, suppose you hold 1,000 "Yes" shares in the "Will OpenAI release consumer-grade hardware products before December 19th?" market, with a market price of US$0.60 per share. If new information emerges that leads the market to believe that the product launch may be later than expected, the price of your shares may drop to $0.39. At this time, your LTV reaches approximately 77% threshold, which will trigger liquidation (and may result in loss of the invested principal).

The design of the liquidation mechanism is the most controversial aspect of this type of agreement. When the borrower's LTV exceeds the liquidation threshold, Gondor will not immediately seize and sell the collateral. Instead, it will first purchase a position with the opposite result in Polymarket to hedge the risk and lock the payment path before disposing of the collateral. This sequence is by design: since the Polymarket order book operates off-chain, seizing collateral first and then hedging will introduce delays and execution risks.

After completing the hedging, Gondor will seize 77% of the collateral, pair up the "yes" and "no" positions, and redeem them at 1 USDC per pair. The remaining 23% of the collateral can be withdrawn by the borrower after the loan is repaid.

Currently Gondor uses a centralized clearing engine and plans to open clearing to external participants after the protocol matures and liquidity improves.

There are several risks associated with this design. Most obviously, Gondor was structurally forced to buy inverse positions when the market was violently repricing. In markets that are rapidly moving or approaching settlement, Polymarket order books are often thin with liquidity and wide spreads, and the bid-ask spread can be extremely wide. If Gondor has to intervene in the reverse market with a market order in this environment, as a predictable liquidity demander, it may lock in unfavorable prices and be sandwiched or front-loaded by more agile robots.

As long as the underlying market in which Gondor provides financing remains relatively shallow, there will be clear incentives for third-party manipulation. Such risks can be expected to be reduced by avoiding illiquid markets and reducing exposure through early closing of positions closer to settlement.

Overall, Gondor clearly demonstrates where the field is headed. The team recently completed a pre-seed round of financing of US$2.5 million, with participation from institutions such as Castle Island Ventures, Maven 11 and Prelude. Its very existence is a bet on the predicted market share that will become the standardized mortgage asset class. For mature traders, Gondor is pushing the prediction market from a gorgeous gambling front-end to a macro derivatives feature.

Space: Leveraged Prediction Market

One of the most important developments in prediction markets is the emergence of leveraged prediction markets. The surge in perpetual contract DEX trading volume has clearly demonstrated the market demand for leveraged financial instruments - perpetual futures contracts have become one of the most successful products in the crypto space. Users can go long or short an asset through leverage, provide a fixed amount of collateral, and be liquidated when the price hits a specific threshold, thereby losing the collateral.

Space, the protocol on Solana, allows users to gain event outcome exposure with up to 10x leverage by simply providing a portion of the position value as collateral. A simple example in its documentation shows how it works:

Market: "Will the U.S. government shut down before the end of the year?"

"Yes" share price: $0.15 (implied probability 15%)

Traders purchase 1,000 "YES" shares - normally $150

When using 5x leverage, only $30 in margin is required (20% of the nominal value).

If the probability rises to 30%, the position value becomes $300, and the $30 margin earns a 500% return on equity (a profit of $150).

If the probability drops to 13.33% ($0.13), the trader will be liquidated and lose $30 in margin.

The leverage mechanism makes prices more sensitive to marginal information. Space also implements a multi-outcome market unification mechanism, where all possible outcomes of the same question (such as who among the five candidates will win the primary) share the same liquidity pool instead of being split into independent contracts. This structure significantly reduces slippage and enables more efficient pricing of complex events.

Think of it as a perpetual contracts exchange for event markets – a direction many market participants expect other platforms to explore. Perhaps Polymarket will eventually support leverage functionality natively in its flagship global platform.

The introduction of leverage significantly enhances the information properties of prediction markets. Leveraged prediction markets speed up the process of integrating new information into prices by allowing traders to express market beliefs with less capital.

At the same time, they also introduce risks common to derivatives markets: liquidation cascades and path-dependent volatility near settlement. In thin liquidity or binary outcome scenarios, these dynamics can amplify market noise as easily as signals. Similar to perpetual contracts trading, the extent to which leveraged prediction markets can improve price discovery will ultimately depend on the depth of liquidity and the adaptability of risk parameters closer to settlement.

Artificial Intelligence and Prediction Markets:

Ethereum founder Vitalik Buterin made several profound insights in his 2024 article "Information Finance", one of which is: Artificial intelligence is expected to greatly expand the feasible design space of prediction markets through high-quality participation in all micro-markets - these markets are originally too small to attract professional human traders.

A typical example is the market on "Midwest Blockchain Conference Research Competition Winners" on Polymarket. The order book depth in this market is extremely thin and almost unprofitable (a $20 buy order increases the probability by about 40%). Human traders have little incentive to price effectively because the opportunity for profit is slim. However, AI agents can evaluate each research proposal at low cost, score its competitiveness, analyze it against the judges’ historical preferences, and continuously update the market price.

The key is that the incentives for running these AI agents need not come solely from trading profits. In many cases, agents may be funded by institutions or individuals who value the production of market information, with trading activities becoming a signal aggregation mechanism rather than an independent source of revenue.

This type of micro market is an area where human attention is scarce and AI attention is abundant. The prediction market here begins to resemble an information engine rather than a gambling product.

AI reduces the cost of participation to close to zero, allowing the market to generate effective signals even if the chips are small. Imagine a scenario where thousands of micro-markets operate in parallel, each efficiently priced by an AI agent.

Interactive interface layer?

Artificial intelligence may also become the interactive interface layer between users and prediction markets. Most people don't know which market to trade, how to size their positions, what relevant markets exist, or whether their idea of ​​an ideal trade is the best way to express their opinion.

In addition, based on the observation of this researcher, the user experience of the Polymarket main platform is quite clumsy - there is often a flaw in which a ribbon animation is displayed to prompt the transaction when the transaction is unsuccessful. This problem has been widely criticized and will be gradually improved in the future through Polymarket's own product iteration and the development of third-party prediction market terminals. In contrast, the user experience of the Polymarket US mobile app (which is reported to use off-chain transactions) is very smooth.

As prediction markets proliferate, users will increasingly rely on AI agents to translate opinions expressed in natural language into optimal on-chain market exposures. Imagine saying to an AI agent: "I think Zcash will break $600 in the next three weeks."

AI does not require the laborious search required by humans:

"Can Zcash break through 1000 US dollars before the end of the year? "

"What price will Zcash reach by 2027?"

"Will Zcash break through 500 before 1231? "

But it can judge instantly:

Is there a Polymarket event contract that directly expresses this view

KalshiWhether the market is pricing the same results more attractively

Are there better risk-reward ratios in other prediction markets

Do leveraged prediction markets (such as Space) provide a purer trading solution

Do perpetual contract long positions on DEX such as Hyperliquid provide better trading options

How to configure positions according to user risk tolerance and capital size

AI can become the user's strategist (scanning relevant markets, assessing liquidity, identifying mispricing and executing trades that best reflect the user's perspective). As the number of markets increases and the granularity becomes finer, human cognitive load will continue to increase. But for AI, this complexity is an advantage rather than a drawback—the more markets there are, the more room there is for intelligent agents to optimize.

The effectiveness of a market depends on the depth of capital that maintains its openness. Without a deep order book, prices will become unreliable and susceptible to manipulation.

Influence Market:

Influence Markets aims to address a fundamental information gap: current prediction markets fail to inform the value of an asset given the occurrence of specific events. This type of information does not exist in a clearly discoverable form. While prediction markets reveal event probabilities and spot markets reveal prices, there is no mechanism to present the market’s collective judgment on asset valuations under specific event-bound conditions—such as how Bitcoin will trade if the Fed cuts interest rates by 75 or 50 basis points, or how Nvidia stock will perform if a candidate with AI threat theory wins an important election.

Influence market users no longer trade probability odds through synthetic yes/no tokens, but directly trade the assets themselves in conditional states, expressing positions such as "I am willing to buy BTC for $110,000 (a 10% premium to the market price) if and only if the Fed cuts interest rates by 75 basis points." This fundamentally enhances the information revealing ability of the market. We no longer need to maintain two independent markets of "Probability of Fed rate cut" and "BTC price" respectively, but directly obtain the "BTC price under the condition of 75 basis points of Fed rate cut" discovery mechanism. This concept can be extended to any asset | event pairing, such as "Google stock price under the condition that GPT-6 is released before Gemini Series 4" or "Gold price under the condition that asteroid mining is realized before 2030."

The key difference is that the event itself and the impact of the event on the company and assets are completely different things. Where prediction markets aggregate the probability of an event occurring, influence markets answer the next question: "What would it mean for the company or asset if this event occurred?" This separation allows each type of market to specialize while creating a more complete set of information.

Decision Market

The decision-making market extends the mechanism of the influence market from information disclosure to governance automation. These markets no longer merely present conditional valuations that inform individual decisions, but directly and bindingly determine whether organizations take action based on outcomes priced higher by the market. Decision markets grew out of Hanson's 2000 working paper "Should We Vote on Values, but Believe?" 》.

This mechanism has been implemented in practice through the "voting DAO", and the total transaction volume of the relevant decision-making market has reached millions of dollars. In a typical setting, an organization makes a decision (such as whether to issue 5% more tokens to fund a new product line), and the market trades two conditional "states": a pass state and a rejection state. Each state sets an independent value for the organization's token, and the token price becomes the market's objective function—that is, what the market seeks to optimize. If the token price is higher in the "passed state", the organization implements the decision; if the token transaction price is higher in the "veto state", the proposal is rejected and no action is taken. Market participants collectively decide which action maximizes expected value, and their trades are executed conditionally based on the final outcome. Galaxy Research has conducted in-depth analysis of these markets and related application organizations in reports on voting systems and their on-chain implementation, as well as in our annual forecasts for 2025 and 2026.

Opinion Market:

The prediction market anchors objective results ("Will X happen?"), while the "opinion market" focuses on subjective issues and narratives. Such markets do not rely on hard external oracles or binary settlement rules, but instead operate as a sentiment barometer.

Many economically relevant issues in crypto, culture, and politics are often difficult to define with clear settlement criteria. For example:

Which Layer2 is currently receiving the most attention?

Is market sentiment more bullish or bearish after Powell’s latest meeting?

These issues are important because they affect capital allocation but cannot be settled by deterministic oracles. Opinion marketplaces like Noise.xyz allow speculation on narratives themselves rather than discrete outcomes.

Opinion markets generally do not settle via deterministic oracles or binary settlement rules. Instead, they operate as continuous sentiment tools, with prices reflecting the market's collective view at a given moment rather than the final "correct" outcome. Traders will make a profit if the price of their bet goes up, and a loss if the price of their bet goes down.

相比之下，预测市场本质上受结算要求约束。要成为可交易标的，问题必须具备客观可验证性、无歧义性，且可由可信数据源判定。例如，可参考Polymarket针对"鲍威尔将在12月新闻发布会中作何表态？"设定的以下规则细则。

4、风险提示

预测市场仍处于相对早期阶段，从"病毒式产品"演变为持久金融基础设施的路径尚未确凿。

当前最大制约因素是流动性。绝大多数市场订单簿深度较浅且价差宽泛，往往仅需数千美元即可使份额价格波动超过10%。在此类市场中，概率数据可能产生误导，市场操纵成为切实隐患。仅最具流动性的市场能提供可靠概率。

理论上，随着市场流动性增强，这一问题将随时间推移得到解决，正如前文所述，AI代理亦有助于有效为小型市场定价。但无论如何，流动性不足是当前存在的现实问题，且将在2026年持续存在。

另一重风险在于市场构建与结算逻辑。预测市场最终依据预设规则结算，若定义存在模糊性或市场标题与结算标准不一致，可能损害信号质量。我们近期在分析Polymarket"美国是否将在……前入侵委内瑞拉"市场时曾论述此问题——该市场因标题与结算定义不符而引发争议。

预言机问题是另一风险点。区块链本身无法感知现实世界，需要外部基础设施将现实信息传递至市场。这些预言机机制的可靠性与治理模式引入了额外风险节点，尤其当特定参与方存在经济激励采取利己行为时。

监管反弹风险仍高悬头顶，尤其在平台向体育以外领域扩张时。即使采用合规模式，若预测市场被定性为鼓励有害行为或"对悲剧事件下注"，仍可能面临政策风险。试想若出现人为操纵预测市场结算而致伤亡的极端案例，随着这类产品日益主流化，吸引此类行为的可能性也将增加。这听似骇人，但我们在pump.fun直播中已目睹过类似为拉抬代币价格而采取的行为。不幸的是，人类确实可能为经济利益采取极端手段。

5、前景展望

预测市场正在从边缘投机产品向基础金融基础设施转型。事件关联合约日益显现出衍生品、对冲工具、抵押资产及信息基础设施的特征。

短期来看，进展将更多由流动性积累驱动，而非新颖的市场设计。尽管增长迅速，多数预测市场流动性依然薄弱，限制了其对成熟资本的实用性，也制约了价格可靠性。那些能成功结合强大零售渠道与清晰监管框架的平台，最有可能提升预测市场作为持久金融工具所需的流动性基准水平。

中期而言，预测市场与传统衍生品的边界将持续模糊。简单的二元合约已在特定场景中发挥类远期对冲功能，而杠杆化预测市场、信用分层及期权型结构正在拓展事件交易的资本效率。尽管这些创新可能加速市场采用（尤其在机构参与者与高净值人群中），它们也将衍生品市场的常见风险引入该领域。

人工智能有望进一步加速这一演进。随着市场数量持续增长且颗粒度不断细化，人类注意力正成为限制因素。 AI代理能够持续为微观市场定价，并跨平台扫描错误定价机会。即使低流动性市场也能通过代理实现低成本高效定价。

观点市场、决策市场与影响力市场或将进一步拓展预测市场的功能边界。这些模式共同推动预测市场从预测事件"是否发生"转向对事件"意味着什么"进行定价。这一转变将拓展其在投资组合构建与决策制定领域的实用性。

展望未来，最可能出现的将是模块化的预测市场生态系统。受监管交易场所、链上协议、封装器、终端、机器人以及人工智能驱动的交互界面，都将推动该领域走向成熟，各自针对不同用户与风险偏好进行优化。

预测市场将逐步融合，最终确立其作为交易不确定性本身的金融基础设施的持久角色。