The Prediction Market Economy
Prediction markets look marginal at less than 0.5% of DeFi volume, but reframed through forecasting, insurance, and hedging they look early rather than small. functionSPACE maps the emerging four-layer prediction economy - infrastructure, venues, builders, and the interaction layer - and argues the next phase of growth won't come from new venues but from where people encounter the ability to express belief.
Ecosystem
By: Igor (@justigor)
Pattern matching is the default mode for all of us. It is a completely human way to interpret something novel, to reach for familiar categories when something new appears. Taking a crypto only frame, prediction markets represent less than 0.5% of total DeFi volume (Messari), and remain dwarfed by the trillions that flow through perpetual futures on an annual basis.
Seen through that lens, they look marginal, and peripheral.
If we re-frame through forecasting, insurance, and derivatives (much as Stablecoins were reframed through payments) the success metric changes. It shifts from "% of DeFi TVL" to "% of global risk, forecasting, and hedging activity." Under that framing, the relevant comparison set is no longer other crypto protocols, but the broader infrastructure that societies already use to manage uncertainty.
Viewed this way, prediction markets appear less like a niche crypto vertical and more like an emerging information-finance primitive.
Our interpretation of the data is that they are not simply another category inside crypto, but one of the clearest examples of a crypto use case showing very early signs of escaping into real-world financial and decision infrastructure.
Lay of the Land
The sector expanded roughly 12× between mid-2024 and early-2025 (Messari), before consolidating into what is now effectively a duopoly, with Kalshi (~42% open interest) and Polymarket (~41%) - although Opinion might have something to say about that.
Data by https://x.com/datadashboards
Against the backdrop of approximately $44B in total 2025 volume (Dune), a broader structural story is beginning to emerge: a four-layer prediction market economy.
What is becoming clear is that the next phase of adoption is unlikely to be driven purely by new venues. Instead, it will be shaped by where and how people encounter the ability to express belief. The interaction layer, rather than the venue layer, increasingly looks like the primary battleground.
The Stack as It Currently Sits
Prediction markets are not a single product category. What is emerging instead is a stack with distinct economics, moats, and innovation surfaces at each level.
Each layer solves a different coordination problem. Each attracts a different kind of builder. Each shapes behaviour in different ways.
An illustrative snapshot, not comprehensive
Layer 1: Infrastructure (Oracles + Primitives)
This layer contains resolution mechanisms such as UMA, Chainlink, and centralised oracles, alongside market primitives like CTF, LMSR, NegRisk, and order books. All of it sits atop settlement rails, token standards, and the broader decentralised web.
Infrastructure succeeds when it becomes invisible. Reliability, credibility, and institutional trust dominate.
For prediction markets in particular, token models that attempt to monetise raw trading volume have historically struggled to sustain long-term alignment.
Shared substrates convert trust into utility. They allow participants to reuse reputational and technical guarantees across contexts. The next major innovation is therefore likely to come from mechanisms that enable new market types and locations, rather than marginally improving existing AMMs.
Instruments shape expressivity, liquidity formation, and incentives at a foundational level. Binary contracts flatten belief and fracture capital. These constraints propagate upward through the entire stack.
Layer 2: Venues
This is the most visible layer. Prediction Index currently tracks more than 140 venues.
Differentiation is attempted through vertical focus, liquidity model variants, creator incentives, social mechanics, and gamification. Despite this experimentation, the vast majority of venues still rely on binary event contracts.
This creates a persistent coordination problem. New venues struggle to bootstrap meaningful depth without subsidies or captive audiences, reinforcing concentration around incumbents.
Coinbase's acquisition of The Clearing Company and its partnership with Kalshi introduces a new institutional-scale challenger. By our estimation, this is more likely to expand overall participation than merely redistribute existing volume, further normalising prediction markets as a financial instrument.
Ecosystem alignment is also becoming central. Base currently dominates venue deployment. BNB and Solana are growing fastest. Monad and Sui are positioning around high-frequency and gamified prediction formats. Venues are increasingly incorporated into the ecosystem value proposition.
Layer 3: Builders
This was the most active layer of 2025.
Builders route users, trades, and liquidity to venues. In practice, they create specialised interaction systems that reframe how markets are discovered, analysed, and used.
Liquidity fragmentation remains the core constraint. Within-market fragmentation is structural. Cross-platform fragmentation is systemic. Liquidity attracts builders. Builders attract liquidity. This feedback loop largely determines where activity concentrates.
Prediction Index tracks more than 112 ecosystem projects, yet only two of the top ten Polymarket builders—BetMoar and OKBet—dominate current volume. Telegram builder groups exceed 300 members each, implying a much larger long tail.
The ecosystem spans solo developers to scaling teams. Its growth trajectory remains exponential.
Terminals and Trading Tools
On Polymarket, builders account for roughly $15M in daily volume (Dune Analytics, 2025). BetMoar alone represents approximately 50% of this.
Most volume concentrates in advanced terminals serving high-value traders, producing heavy-tailed usage distributions. Many tools are optimised for professional workflows rather than mass adoption.
Some newer interfaces, such as Polymtrade with ~650 DAU, demonstrate that smaller products can still generate meaningful throughput. Projects like Quantish point toward specialised agents and domain-specific interfaces, signalling that the next phase of innovation will involve deeper integration with user context rather than further refinement of generic terminals.
Builder Economics
The commercial model remains unsettled.
Polymarket has been explicity about future fee introduction, with 15-minute markets already incorporating fees. Kalshi's DFlow stack introduces multi-hop extraction. Across both, most current revenue flows through leaderboard incentives and prize pools.
Builders therefore face persistent platform risk. Fee changes or native feature replication can quickly erode independent products.
Bonna Zhu's Nothing Research analysis observed that exchanges retain control over depth and discovery, placing downstream builders in structurally dependent positions. Sustainable projects increasingly focus on defensibility through data, workflow, community, or domain expertise.
At the same time, AI tooling is dramatically lowering barriers to entry. Without advances in liquidity and resolution primitives, this will intensify competition without expanding system capacity.
Layer 4: Apps and Interaction
This layer embeds prediction inside existing workflows rather than presenting it as a destination.
Robinhood's Kalshi integration now accounts for more than 50% of Kalshi's volume, particularly in sports. Coinbase's deployment is likely to show similar dynamics.
These platforms aim to become comprehensive financial environments. Prediction is incorporated as a feature, not as a separate product.
Because these integrations rely on Kalshi-hosted contracts, expressivity remains limited. Liquidity is not a barrier. Regulatory speed is however.
A neutral substrate would allow large applications to design proprietary economic models and support markets unavailable under centralised structures.
DeFi, Wallets, and Distribution
Polymarket, Kalshi, and Myriad have integrated into major wallets and DEX interfaces, driving sustained growth of approximately 5–10% month-on-month.
Direct monetisation remains weak due to low fees, low volatility, and interface friction. Nevertheless, embedding markets where users already transact remains structurally sound. Lowering experimentation cost toward neutrality will accelerate this trend.
Snapshot of Jupiter volume - @datadashboards
Beyond Crypto: The Broader Context
In crypto terms, PMs are small; in "information finance" terms, they are early instruments for trading uncertainty. Prediction markets are converging toward information infrastructure.
They increasingly compete with polling, expert aggregation, decision-support systems, and risk analytics. The "culture markets" thesis advanced by commentators such as 0xWeiler highlights attention-driven embedding as a major 2026 vector.
Mechanism innovation remains under-explored. Numerical-range prediction systems operating on shared liquidity surfaces face little direct competition. Binaries fragment belief and capital, while continuous distributions preserve informational density.
This enables new forms of expression even in entrenched domains such as sports, macroeconomics, crypto, earnings, and weather.
Interaction is the distribution
"Prediction markets will converge toward a durable role as the financial infrastructure for trading uncertainty itself." — Matt Huang of Paradigm, via Galaxy Research
The next order-of-magnitude expansion comes from embedded predictions appearing where decisions occur. Venues remain important, but horizontal scaling alone cannot capture the broader forecasting market.
The addressable market includes enterprise risk management, supply chain forecasting, insurance underwriting, portfolio hedging, and policy simulation. McKinsey estimates the global analytics and forecasting market at over $70B annually and growing.
Uncertainty is encountered in context: while reading news, allocating capital, planning operations, or managing exposure. Prediction surfaces embedded inside these workflows reduce friction and increase participation. An early example of this is Forbes embedding forecasting surfaces into their news site.
Sports: cadence + clarity + existing fan attention. Sports keeps acting as the liquidity magnet.
Culture: fastest potential growth in some datasets because there are fewer non-PM substitutes for trading cultural outcomes.
Crypto and economy: growth driven by faster-resolving, repeatable markets rather than long-cycle political parking.
Robinhood provides the early signal that demand exists when expression is contextual.
If structural and usability constraints are addressed, prediction markets become general-purpose belief infrastructure. Beliefs evolve into financial primitives.
The "Fee-Less" Misconception
Many participants assume trading mechanisms require tolls. Yet when trust is the scarce resource, monetisation can occur upstream or downstream. Cost-neutral primitives increase experimentation, surface diversity, and reduce extractive pressure.
Infrastructure becomes not only reliable, but credible.
What becomes possible when belief formation is no longer metered at the base layer?
From Destinations to Rails
The abstraction layer makes PMs feel native in context. Lower creation costs and neutral resolution change the economics for builders and apps.
Belief lives across interfaces, agents, and applications. That transition accelerates the growth of the prediction economy.
References
https://www.galaxy.com/insights/research/prediction-markets-impact-markets-decision-markets-futarchy
https://pond.dflow.net/learn/prediction-market-fees#prediction-market-fees-and-rebates
Igor leads research at @functionspaceHQ an open-source project exploring market-led resolution and novel economic instruments for prediction markets.
More Research
Explore our additional research for more in-depth insights.
