Written by: Nishil Jain
Translated by: Block unicorn
Introduction
I often open the Polymarket website to check the latest developments on global hot news. For example, how likely is it that the U.S. will occupy Greenland? Or, what will be the final valuation (FDV) of Monad at launch?
Vitalik is also using the market as a news tool, and perhaps many other analysts and individuals around the world have made this shift as well. Soon, we will see these markets integrated into our existing news feeds and social applications.
In fact, we have yet to hear any updates on last year's collaboration between X and Polymarket. Just last October, as part of its strategic investment of up to $2 billion, Intercontinental Exchange (ICE, the parent company of the New York Stock Exchange) announced plans to explore the commercialization of Polymarket data.
While these information flow integrations are still ongoing, prediction markets have successfully entered financial applications like Coinbase and Robinhood. Kalshi is making its way into the web3 application market through the Jupiter super app and Phantom wallet, while Polymarket has established partnerships with MetaMask and Rabby Wallet.
The trading volume of prediction markets remains strong even after the peak of the elections. Currently, the daily trading volume of prediction markets has exceeded $800 million—at the time of writing, surpassing Pumpfun's $775 million trading volume.
But today, we are not here to discuss the obvious phenomenon of the rise of prediction markets (which is now a recognized fact). We will explore the upcoming development trends.
The rise of pump.fun has spawned an ecosystem composed of Telegram bots, trading terminals, and DeFi ecosystems that support these tokens. Similarly, we are now seeing the emergence of some applications that extend the concept of prediction markets beyond existing forms—such as lending against prediction market positions, leveraging your predictions, opinion markets, decision markets, and more.
I know all of this sounds strange, but by the end of this article, you will understand everything.
Galaxy Research recently published an article titled "The Future Landscape of Prediction Markets," which elaborates on emerging mechanisms that go beyond basic event predictions and their technical details. This article draws on the concepts discussed in that article and adds some personal insights.
Overall, next-generation prediction market applications are developing in three main directions:
Expanding into decentralized finance (DeFi) through derivatives and lending,
Applying artificial intelligence to predict outcomes and becoming the default interface for interacting with results, and
The evolution of prediction markets themselves, going beyond mere event predictions to include decision markets, opinion-based markets, and markets predicting the impact on asset prices.
DeFi Expansion
When a platform gathers a crowd with varying risk appetites and trading habits, it becomes quite challenging to meet everyone's needs on a single platform. For example, users wanting to exchange ETH for USDT cannot do so on Uniswap. This is where Aave or Hyperliquid comes into play.
Gondor is building a lending platform for prediction market traders. Once funds are deposited, traders can borrow up to 50% of USDC against their holdings as collateral. The borrowed funds will be returned directly to Polymarket and displayed as the trader's "cash balance."
Through Gondor, traders can free up previously idle funds while maintaining their investment exposure and reinvesting the funds into new trades.
Of course, there are corresponding rules. The Gondor team manually screens specific market positions that meet collateral requirements. These rules ensure that illiquid and easily manipulated markets are excluded from the platform. Many factors determine which markets are whitelisted: the depth of the order book, the clarity of solution standards, and the completion time of solutions.
The team recently raised $2.5 million in seed funding, with investments from Castle Island Ventures, Maven 11, and Prelude. The purpose of Gondor is to bet that prediction market shares will become a standardized collateral asset class.
Then there is Space, which allows users to gain leveraged exposure to event outcomes. Suppose there is a market on "Will Monad FDV's market cap exceed $8 billion at launch?" where the price for "yes" is $0.15 (implying a 15% probability). A trader buys 1,000 shares of "yes"—typically this would require $150, but with 5x leverage, only $30 in margin is needed. If the probability rises to 30%, that position will be worth $300, yielding a 500% return on equity (a $150 profit). If the probability drops to 13.33% ($0.13), the trader will be liquidated, losing the $30 margin.
Space allows traders to profit even from slight changes in probability. For example, those expecting an overall rise in the cryptocurrency market can bet on the Monad FDV market and profit from its slight increase in probability.
Moreover, there are many cryptocurrency users from the meme coin ecosystem, AI coin projects, etc., who trade with the goal of turning $10 into $1,000. For them, as well as those with extensive market knowledge but limited funds, these markets may represent an attractive investment channel.
AI and Prediction Markets
Vitalik has shared some insights on the application of artificial intelligence in prediction markets on his blog "Info Finance."
One of his points is to leverage AI to achieve high-quality participation in all micro-markets; otherwise, these micro-markets are too small to attract skilled human traders.
Many of the most interesting prediction markets can help us gather information, targeting "micro" questions: millions of micro-markets, each with relatively small decision consequences.
Importantly, the motivation for operating these agents does not solely come from trading profits. In many cases, the funding for the agents may come from organizations or individuals that value market information output, where trading activity serves merely as a signal aggregation mechanism rather than an independent source of income.
In these micro-markets, human attention is scarce, while AI attention is abundant. Here, prediction markets are no longer seen as gambling products but rather as information engines.
AI reduces participation costs to nearly zero, allowing markets to generate meaningful signals even with small trading volumes. Imagine thousands of small markets operating in parallel, each efficiently priced by AI agents.
The second use of AI is as the default interface layer for prediction markets. As the number of markets increases, tracking these markets across multiple platforms becomes a significant cognitive burden for any trader. AI can convert users' natural language inputs into tradable positions. For example, a user can tell the AI that they believe silver prices will break $100 per ounce by the end of February—the AI can then find potential trading opportunities for the user to profit from this insight—whether through silver options or silver price prediction markets.
Evolving Prediction Market Design
Impact Markets
There is a strange gap in how today's markets operate: you can bet on whether something will happen, and you can see the current price of an asset, but it is challenging to discover what the market believes the value of that asset will be if a specific event occurs.
Prediction markets show that there is a 60% chance Trump will raise tariffs on China. The spot market may show that Bitcoin is trading at $100,000, but you cannot know what the market generally believes the trading price of Bitcoin will be if the tariffs are indeed implemented.
Impact Markets solve this problem. Instead of betting on a "yes/no" basis through synthetic tokens for events, it trades actual assets with conditional states. You are essentially saying, "I will only be willing to buy Bitcoin at $90,000 (10% lower than the current market price) if the tariffs are imposed."
This applies to any asset-event combination. "If Robinhood launches cryptocurrency trading, use COIN as the underlying," "If asteroid mining becomes a reality by 2030, use gold as the underlying." You get my point.
Events and their impacts are fundamentally different. Prediction markets answer "Will this happen?" while impact markets answer "What will happen to this asset if this event occurs?"
Decision Markets
Decision markets take it a step further, moving from information transmission to action. They not only reveal market thoughts but also allow the market to determine the actual actions of organizations.
This functionality has already been implemented through platforms like MetaDAO with Futarchy DAO. Here’s how it works:
An organization proposes a decision, such as firing its Chief Technology Officer (CTO). The market trades two conditional states: Pass and Fail. Each state will have different pricing for the organization's tokens. If the token price rises in the Pass state, the proposal is executed; if the token price rises in the Fail state, the proposal is rejected.
The market determines which action maximizes expected value. Trades are settled only based on the winning outcome.
Opinion Markets
Not all important questions have clear answers. Prediction markets require objective results—"Did X happen?" These results are easily verifiable. But many economically related questions are not so straightforward.
Which prediction market is currently the most watched? Is market sentiment more optimistic after Powell's last speech? These questions affect capital allocation but cannot be simply answered with "yes" or "no." This is where opinion markets come into play.
Platforms like Noise.xyz allow you to speculate on various narratives themselves. These narratives are not determined by deterministic prophecies or binary rules but serve as ongoing sentiment tools. Prices reflect the market's collective view at any given moment rather than a final "correct" outcome.
If the price of the opinion you bet on rises, you make money; if the price falls, you lose money. It's that simple.
Traditional prediction markets are limited by adjudication requirements: questions must be objectively verifiable, unambiguous, and answerable through credible data. Opinion markets trade in sentiment, which often changes continuously.
Questions to Consider
Prediction markets are still in their early stages, and most markets have low liquidity. A few thousand dollars can cause share prices to fluctuate by more than 10%. In this case, the probabilities displayed on the screen are less about "the market's general view" and more about "the view of the last person who put in $5,000."
Only the most liquid markets, such as presidential elections or major sporting events, can provide reliable price signals. Everything else is susceptible to manipulation. Those with sufficient funds and strong beliefs can artificially create consensus. Once involved, they can achieve this through persistent price fluctuations in some lesser-known markets.
Oracles are another potential fault point. Blockchains cannot perceive the state of the real world. They require external infrastructure to input information. Who controls this infrastructure? How do they make decisions? What happens when the flow of funds depends on their judgment?
This issue arose in the Polymarket trade regarding "Zelensky wearing a suit." Despite multiple news outlets reporting that he wore a suit, the market authority determined that he did not, leading to intense controversy. The crux of the dispute was the style of the suit he wore, which clearly has no definitive answer.
Those running the oracle have economic incentives. Sometimes, these incentives drive them to provide accurate information. Other times, they do not. Or most of the time, the information itself may be ambiguous.
As these markets develop and mainstream, we will see transformations, and platforms will mature to address these issues.
Let's check back on their situation in a few months.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
