Prediction markets are no longer a niche experiment. In 2025, the sector processed $63.5 billion in trading volume — a 302% year-over-year increase. Polymarket and Kalshi now capture 85–90% of global volume, ICE (Intercontinental Exchange) has made a $2 billion strategic investment in Polymarket,...
Prediction markets are no longer a niche experiment. In 2025, the sector processed $63.5 billion in trading volume — a 302% year-over-year increase. Polymarket and Kalshi now capture 85–90% of global volume, ICE (Intercontinental Exchange) has made a $2 billion strategic investment in Polymarket, the CFTC has granted full regulatory approval for U.S. operations, and combined industry volume is on track to exceed $100 billion in 2026. Wall Street has arrived.
But even as institutions pile in, Ethereum co-founder Vitalik Buterin — an early Polymarket investor — is sounding the alarm. In a February 14 post on X, he warned that prediction markets are "over-converging to an unhealthy product market fit" dominated by short-term crypto bets and sports gambling. His alternative vision is radical: transform prediction markets into AI-powered personalized hedging engines that could, in his words, "replace fiat currency" entirely. The proposal arrived just days before his ETHDenver 2026 keynote, where he expanded on how Ethereum could serve as the economic coordination layer for AI agents in a world of "perfect markets."
This report examines the collision between Wall Street's institutional buildout of prediction markets as a new asset class and Buterin's philosophical counter-proposal to turn them into the foundation of a post-fiat financial system.
The numbers tell a story of explosive growth. In 2025, Kalshi cleared $43.1 billion in total volume while Polymarket recorded $33.4 billion, combining for over $76 billion between the two dominant platforms alone. The broader market hit $63.5 billion in tracked volume — a 302% increase from the prior year.
As of February 20, 2026, daily volumes paint an evenly matched duopoly: Kalshi at $467.1 million (54% share) versus Polymarket at $401.8 million (46%). On a 30-day basis, Polymarket leads with $7.6 billion (+34.1%) while Kalshi is closing fast at $6.5 billion (+60.5%). At current run rates, the industry is on pace to blow past $100 billion in 2026 volume.
The composition of that volume has shifted dramatically. Sports betting grew from 5% of Polymarket's volume to over 40% by early 2026. Crypto-related markets now account for more than 25% of total volume. Political markets — the category that put prediction markets on the map during the 2024 U.S. election cycle — have been supplanted as the primary driver. Per-meeting volume on political contracts alone grew from $59 million in September 2024 to $660 million by January 2026, an eleven-fold increase.
Yet this growth is happening against a brutal macro backdrop. The crypto market is deep in a bear cycle: Bitcoin has fallen from above $80,000 to the low $60,000s, Ethereum trades around $1,900–2,000, and crypto ETFs have posted four consecutive weeks of outflows totaling $3.8 billion, pushing total AUM to $133 billion — the weakest since April 2025. Prediction markets, paradoxically, are booming while the underlying crypto markets contract.
The institutional capture of prediction markets accelerated through a series of moves in late 2025 and early 2026 that would have been unthinkable two years prior.
ICE's $2 Billion Bet. Intercontinental Exchange — the parent company of the New York Stock Exchange — finalized a $2 billion strategic investment in Polymarket in October 2025, valuing the platform at $9 billion. ICE subsequently launched "Polymarket Signals and Sentiment," a tool that delivers normalized prediction market data feeds to professional traders. ICE is now the exclusive provider of Polymarket data for institutional capital markets, embedding prediction market probabilities into the same infrastructure used by hedge funds and bank trading desks.
CFTC Regulatory Approval. In November 2025, the CFTC issued an Amended Order of Designation granting Polymarket permission to operate as a federally regulated exchange in the United States. The platform had bypassed the traditional multi-year registration process by acquiring QCEX, an already-registered platform, for $112 million. Under the new framework, Polymarket must implement enhanced surveillance, clearing procedures, and full Part 16 reporting. Users can now trade through Futures Commission Merchants (FCMs), integrating prediction markets into the same regulatory infrastructure as derivatives markets.
CFTC Chairman's Stance. Chairman Mike Selig has declared that prediction markets are not gambling and that the CFTC has "exclusive jurisdiction" to regulate the space — a critical legal distinction that shields the industry from state gaming commissions and opens the door for mainstream financial integration.
Polymarket relaunched for U.S. users in December 2025 under stricter compliance requirements including mandatory KYC and regulated intermediaries. Nearly four years after being shut down by the CFTC, the platform's return marks the definitive mainstreaming of "information finance."
Into this triumphant institutional moment, Vitalik Buterin dropped a grenade.
On February 14, 2026, the Ethereum co-founder — himself an early Polymarket investor — published a post on X warning that prediction markets are drifting toward what he called "corposlop." He described the current trajectory as platforms "over-converging to an unhealthy product market fit: embracing short-term cryptocurrency price bets, sports betting, and other similar things that have dopamine value but not any kind of long-term fulfillment or societal information value."
Buterin acknowledged the economic mechanics at play: "There is nothing fundamentally morally wrong with taking money from people with dumb opinions," he wrote, but argued that overreliance on this dynamic creates perverse incentives for platform operators. If the business model depends on a steady supply of uninformed bettors making poor wagers, platforms are incentivized to attract — not educate — users.
This represents a notable shift from his December 2025 position, when he described prediction market participation as "healthier" than traditional market speculation. In two months, Buterin went from cheerleader to critic, urging the industry to "build the next generation of finance, not corposlop."
The critique is not merely philosophical. It carries economic weight. If prediction markets settle into a sports-betting equilibrium — high volume, low information value — they become regulatory targets (state gaming laws) rather than financial infrastructure (CFTC-regulated markets). The distinction between a "market for information" and a "gambling platform" is precisely the legal argument that enabled Polymarket's return to the U.S. Losing that narrative could undo years of regulatory progress.
Buterin's alternative vision is both technically ambitious and economically radical.
The Core Mechanism. Instead of betting on binary outcomes (Will Bitcoin hit $100K? Will the Chiefs win the Super Bowl?), Buterin proposes prediction markets indexed to categories of goods and services — food, housing, transportation, healthcare — separated by region and demographic profile. These markets would function as hedging instruments, allowing participants to offset future cost increases.
The AI Layer. Each user would run a local large language model (LLM) that analyzes their personal spending patterns and regional economic conditions. The LLM would construct a personalized basket of prediction market positions representing "N days of that user's expected future expenses." When a user wants price stability, they hold these positions. When they want growth, they hold ETH, stocks, or other yield-bearing assets.
The Insurance Analogy. Unlike speculative betting — where one party wins and another loses — hedging creates mutual benefit. Both counterparties reduce their risk exposure. Buterin's example: a biotech shareholder could take positions in election prediction markets to hedge against political outcomes threatening the sector, improving risk-adjusted stability without seeking speculative profit.
Yield-Bearing Denomination. Critically, Buterin proposes denominating these markets in yield-generating assets — interest-bearing instruments, wrapped equities, or ETH — rather than non-interest-bearing fiat stablecoins. This solves a fundamental incentive problem: if hedging positions are denominated in productive assets, both sides of the trade earn yield while managing risk. Non-yielding fiat creates an excessive opportunity cost that makes the structure uneconomical.
The endgame of Buterin's proposal is the elimination of fiat currency as a unit of daily economic life.
The logic runs as follows: today, people hold fiat currency (or fiat-pegged stablecoins) because they need price stability for everyday transactions. But fiat is a blunt instrument — it provides uniform stability regardless of individual spending patterns. A software engineer in San Francisco and a farmer in Iowa have radically different cost structures, yet both are served by the same dollar.
Buterin's system would replace this one-size-fits-all stability with personalized hedging portfolios. Each individual's AI assistant would maintain a dynamically adjusted basket of prediction market positions calibrated to their specific expenses. The result: "People can hold stocks, ETH, or whatever else to grow wealth, and personalized prediction market shares when they want stability."
If this sounds like a thought experiment, the infrastructure to support it is closer than most realize. Polymarket already processes nearly $500 million daily. CFTC-regulated intermediaries exist. ICE is distributing prediction market data to institutional traders. The missing pieces are the specialized commodity indices, the AI personalization layer, and the regulatory framework for hedging-oriented prediction markets.
The economic implications are staggering. If prediction markets can deliver personalized price stability without fiat, the demand for non-interest-bearing money — including stablecoins — diminishes. Central bank monetary policy loses a transmission mechanism. The entire concept of "inflation" becomes personalized rather than aggregated. CPI, as a policy tool, becomes an anachronism.
Buterin's prediction market thesis was not delivered in isolation. It arrived days before ETHDenver 2026, where he delivered a keynote titled "The Next Epoch of Ethereum" to over 25,000 attendees at Denver's National Western Center.
In the keynote, Buterin revisited the idea of "perfect markets" — systems where information flows freely, barriers to entry are low, and participants transact without dominant intermediaries. He laid out a four-pillar framework for Ethereum's role in an AI-driven economy:
The connection to prediction markets is direct: if AI agents are going to manage personalized hedging baskets on behalf of users, they need an economic coordination layer with trustless settlement, privacy, and programmable payments. Ethereum, in Buterin's framework, is that layer. The prediction market vision and the AI agent vision are two sides of the same coin — a financial system where software agents manage personalized economic risk on decentralized infrastructure.
This philosophy aligns with Buterin's broader "d/acc" (defensive acceleration) framework, which advocates using technology to strengthen decentralized cooperation rather than concentrating power. In this view, AI-powered prediction markets are not about efficiency for efficiency's sake — they are about giving individuals sovereign control over their economic exposure.
The prediction market industry is living through a paradox. It has never been more institutionally legitimate — ICE data feeds, CFTC regulation, $100 billion volume trajectory — yet its most prominent philosophical champion is warning it is on the wrong path. Buterin's intervention matters not because he can redirect $100 billion in flows overnight, but because he is articulating the economic logic that originally justified prediction markets as something more than gambling platforms.
The hedging thesis — personalized, AI-driven, denominated in productive assets — is technically feasible but economically untested. No one has built granular commodity prediction markets at scale. No one has deployed LLM-driven personalized hedging baskets. The regulatory framework for hedging-oriented (rather than speculative) prediction markets does not yet exist.
But the directional bet is clear. If prediction markets remain sports-betting platforms with blockchain settlement, they are a $100 billion niche. If they evolve into personalized economic hedging infrastructure, they are a potential replacement for the $7 trillion daily foreign exchange market and the $130 trillion global money supply. Buterin is pointing at the bigger prize — and daring the industry to pursue it.