Prediction Markets: SocialFi and TAM

Related Writing:

• The Future Of Ownyfans (The token is the product)
• RWA vs NWA (Looking for token-culture fit)
• Betting On 7 Deadly Sins (Broader SociaFi thesis and hyperspeculation)

This essay, a standalone piece, partly revisits our thinking over the past two years, progressing from abstract concepts to specific observations. While it could be considered a loose sequel to previous writings, prior familiarity is not required.

Our exploration began in mid-2023 with the rise of the now-infamous (or perhaps forgotten) friend.tech. Today, we identify two dominant trends within what we broadly categorize as SocialFi:

1. Memecoins
2. Prediction markets

Memecoin mania saw tokens, often spun out on pump.fun, existing freely on Solana, while shilled across platforms like X and TikTok. It was not a neatly gated and comprehensive social in-app experience. It was a cross-platform memetic game of fluctuating prices, where volatility was the norm and rugs were a feature, not a bug.

Prediction markets, particularly Polymarket, gained significant traction in the summer of 2024, driven by the intensifying US presidential election. Evolving from a niche information source, Polymarket entered the mainstream. Despite their cyclical nature, largely tied to the US election, prediction markets remain a widely discussed topic with a long-term upward trend in usage.

To better understand these two trends, we frame their underlying mechanisms as opposites on the SocialFi spectrum. Both involve market launches, but one is attention-driven, while the other is event-based.

The reason to frame these two concepts as opposites is simple. One involves a bounded game where price is the news and the spectacle itself, whereas the other relies on an exogenous event to settle the price. Also, it seems that founders try to build everything in between, and so far, they are failing in capturing more users than either of these two opposites.

Both concepts are simple, whereas everything in the middle appears convoluted and overengineered. Their simplicity and the fact that these markets are not tied to a specific platform, even if launched on one, may contribute to their widespread adoption across various platforms and channels.

Memecoins have been extensively discussed, and their attention-market mechanics are straightforward. They represent a new form of media where the token is the product and the price is the news. While simple and unsophisticated, they offer entertainment value.

In a similar fashion, we'd like to explore prediction markets in a wider context to understand their full potential. It's also interesting to consider if they are more than just a novel form of gambling, and whether hyperfinancialization is distinct from, or synonymous with, hyperspeculation.

It is important to note that relying on hyperspeculation as a thesis is rather overplayed, and as a standalone thesis, it does not hold well. Especially after mediocre minds have adopted it as a dogma. It’s not a secret and without a nuance the thesis of hyperfinancialization is worth nothing.

The Rise Of Prediction Markets

Having been in the industry long enough to remember Augur and the original Gnosis, we recall that prediction markets were a frequent fascination of the crypto community. Closely tied to Robin Hanson’s concept of futarchy, the 2017 wave of crypto projects ambitiously aimed to overhaul how society governs itself (not only online).

Futarchy is a theoretical framework for governance via prediction markets, essentially choosing policies based on market votes. Despite its philosophical value, it hasn't been widely adopted beyond a primitive form of granular and isolated event-driven prediction markets.

While initial crypto prediction market experiments involved tokens and, for that time period, perhaps an overengineered design, commingling reputation and financial capital. The real breakthrough was Polymarket. This simple product, built on Polygon with a permissioned market launch system, stripped away any token mechanics from the user experience and relied on a questionable oracle solution.

Products needn't be perfect to acquire attention and liquidity. Polymarket slowly but surely became a dominant prediction market, often cited by mainstream media and pop icons like Elon Musk. As mentioned above, Polymarket entered a period of rapid growth with the 2024 US election heating up, with the traction peaking around the election and inauguration of Donald Trump.

Source: https://dune.com/filarm/polymarket-activity

During Polymarket’s rapid growth, we noticed a significant increase in our pipeline of teams building prediction markets, prediction market-adjacent products, or memecoin-prediction market hybrids. To this day, however, the only relevant competitor to Polymarket is Kalshi.

Kalshi's core platform is not onchain; it's a centralized exchange operating under US regulatory oversight. However, in September 2025, Kalshi launched KalshiEco, a blockchain-integrated ecosystem hub in partnership with Solana and Base to expand into onchain prediction markets.

By September 2025, Kalshi captured over 60% of the prediction market volume (with sports as a main focus), surpassing competitors including Polymarket. In the meantime, Polymarket can now operate in the US after receiving clearance from the CFTC in early September 2025, ending a three-year ban that previously blocked American users.

With the two competitors planning an aggressive expansion, the questions that we ask are:

1. Are prediction markets a product or a feature
2. How big is the TAM for prediction markets?
3. Do these markets need to be onchain?

A Product Or A Feature

To further clarify, we are exploring whether prediction markets function best as a standalone application or as a feature integrated into a larger social media platform like X or into a more robust financial services provider like Robinhood.

Prediction markets, similar to pump.fun and memecoins, lack bundling in their offerings. A social media platform, such as X, could integrate these markets directly into its newsfeed. Given that social media thrives on interactive events and spectacles, embedding a financial layer into an existing network and feed would be a feasible undertaking.

Are prediction markets betting or can they be thought of as investing? Either way, they facilitate speculation on future events. This could be bundled into existing financial services applications. Robinhood already offers “event contracts”, enabling users to trade on the outcomes of real-world events like sports games, elections, economic indicators, culture, and more.

To answer a product or feature dilemma, we must understand whether users will favor the convenience of vertical integration or the freedom of ad-hoc market creation. Will the future be marked by an ever-increasing force of centralization, or will the pendulum swing back?

Both startups and established companies aim to build a mythical superapp, exemplified by WeChat's aggregation of social and financial features. For any standalone feature, such as a prediction market, creating a moat and withstanding pressure from aggregators (superapps) could result in leaning on token-incentivized usage. This, however, would be a departure from established best practices set by exchanges, stablecoin issuers, etc. that created their moats without the token.

Prediction Market TAM

The potential of prediction markets lies in their ability to serve as the best financial expression for event-driven outcomes. While a portion of this market could stem from existing betting markets, such as sports, which represent a $100 billion industry, the true innovation lies in the creation of entirely new markets.

This raises several questions: Does the granular, event-driven financialization of information through prediction markets truly appeal to market participants? Do these users desire more markets to bet on? Can prediction markets effectively distinguish signal from noise amidst an abundance of information?

In a Hayekian sense, the market functions as a manifestation of a knowledge graph, where price abstracts the essential information. If prediction markets expand access to more markets for more people in more ways, the knowledge graph will grow, consequently increasing the volume of information traded.

We are witnessing a convergence of abundant capital (fiat money) and abundant information, hinting at an infinite total addressable market (TAM). While this may be an exaggeration, it wouldn't be surprising to see layers of markets emerge, with users making granular bets on company earnings, finding prediction markets a superior method to express their views compared to traditional stock trading.

This outcome is not inevitable, but one might consider whether financial derivatives were historically inevitable and necessary. They may have initially served only to express a particular bet more accurately (or granularly) than non-derivative trading, yet they ultimately grew to surpass the size of the spot market itself as they became a tool for leverage and created exotic markets despite the current lack of convex payoffs.

Onchain Future

As crypto investors, we are naturally inclined to advocate for a decentralized, onchain future for prediction markets. While we acknowledge that these markets can and do exist as centralized services, we will outline our rationale for why on-chain solutions will ultimately prove superior and achieve long-term success.

We believe that permissionless creation and access to markets offer a more robust foundation for building the financial infrastructure of the future. The full value of this approach may not be immediately apparent, but establishing and resolving these markets on a credibly neutral layer ensures an unbiased and transparent process, even as oracle designs continue to improve¹.

---

¹Some argue that user-created markets will have terrible resolution because of poorly-defined criteria. However, similar concerns existed for amateur content creators in the pre-YouTube era. This period saw the rise of independent content creators producing high-quality content. Similarly, permissionless market creation could lead to the emergence of a new class of "power users" who deliver high-quality markets

Launching these markets onchain, and allowing them to be untethered, similar to memecoins, could foster greater interactivity compared to a constrained in-app experience. Just as any token can be traded across multiple decentralized and centralized platforms, so too could these markets. Onchain deployment facilitates seamless embedding across various providers.

The inherent onchain nature of these markets could enable easier composability of individual markets, allowing for further bundling or even potential rehypothecation. While we are not advocating for financial shenanigans, the market's nature tends to favor the adoption of sophisticated or complex financial products.

Being onchain enables novel design mechanisms e.g. different approaches to markets with oracle-based pricing tradeoff, enabling complex products that simple CLOB products cannot have (parlays, system bets, 10x more different market types), leveraging Pool-vs-Peer AMM just like Overtime does. This is similar to how Uniswap disrupted order book exchanges.

Onchain environments foster greater asset fluidity. In our view, prediction markets, as a novel financial instrument, maximize their potential when launched in a transparent, permissionless, and composable manner. Furthermore, we believe there is an intriguing design space for a token that governs a prediction market launcher.

A token linked to a prediction market (launcher) could establish a moat, enhance the security and reliability of markets, and help bootstrap long-term liquidity. Such a token is expected to strengthen the network effects of a given launcher. After all, similar to how pump.fun serves as a case study for memecoins, individual markets (as in events) are essentially memes whose resolution depends not on the price itself, but on an exogenous event.

Still, the token case studies tell us more about failures rather than successes. That does not mean there is no future iteration that could yield a superb user experience. If onchain markets have any future, the token design should still be considered in its infancy, and future experimentation is a must.

Summary & Concluding Thoughts

The question of whether prediction markets function as a product or a feature is central to understanding their long-term trajectory. While the convenience of vertical integration within existing platforms like X or Robinhood is a compelling argument for their adoption as a feature, the freedom and innovation offered by ad-hoc market creation as a standalone product cannot be overlooked.

The success of either approach will largely depend on user preferences for bundled services versus independence of onchain launchers, and the ability of prediction markets to cultivate strong network effects and potentially leverage token incentives to establish a defensible moat.

The total addressable market (TAM) for prediction markets appears vast, extending beyond traditional betting into granular, event-driven financialization of information. This expansion could lead to the creation of entirely new market categories, allowing for more precise expressions of views on a multitude of future events.

Framing the market as an ever-expanding knowledge graph, where price abstracts essential information, suggests that prediction markets could grow aggressively as a new market layer. This growth mirrors the historical expansion of financial derivatives, which eventually surpassed the size of their underlying spot markets.

Ultimately, the onchain future of prediction markets, while not universally adopted, presents a compelling vision for their long-term success. The permissionless creation and access, credible neutrality, and composability inherent in onchain solutions offer a robust foundation for building future financial infrastructure.

This environment fosters greater asset fluidity and opens up intriguing design spaces for token-governed prediction market launchers, which could establish moats, enhance security, and bootstrap liquidity, ultimately treating individual markets as event-dependent memes within a decentralized ecosystem.

The above still does not refute the possibility of the market being monopolized by either a centralized or decentralized solution. On the contrary, we believe that the market will follow a long-tail distribution.

We advocate for continued experimentation and independent thinking in the bottom-up design of SocialFi applications, particularly those that harness the memetic nature of today's markets. Key features still to be explored include shorter-duration markets, a shift from CLOBs to bonding curve innovations, and the integration of these with permissionless market creation.

Disclaimer: Zee Prime invested in Polymarket and Overtime.