- Prediction markets are exchanges where people trade contracts tied to real-world outcomes, and the resulting prices function as crowd-sourced probability estimates.
- Polymarket processed over $1 billion in Iran war bets this month, reigniting debate over whether betting on geopolitics is journalism, gambling, or market intelligence.
- The NBA, Congress, and the CFTC are all wrestling with prediction markets at once—some want to ban them, others want to use them.
Prediction markets aren’t new. The Iowa Electronic Markets have tracked US presidential elections since 1988, and academic research consistently shows that market-generated probabilities outperform both polls and individual expert forecasts. But 2026 is the year prediction markets went mainstream—and the year everyone started arguing about what they actually are.
Polymarket, the largest crypto-native prediction platform, processed over $1 billion in bets tied to a potential US-Iran conflict this month. The NBA signed a data partnership with a prediction market operator. Congress introduced bipartisan legislation to regulate them. And the CFTC—America’s commodity watchdog—is suing some platforms while quietly considering how to license others.
Here’s how prediction markets work, why they’re suddenly everywhere, and what the regulatory fight means for the future of crowd-sourced forecasting.
What Is a Prediction Market?
A prediction market is an exchange where people buy and sell contracts that pay out based on real-world events. Each contract represents a specific outcome: “Will Bitcoin hit $100,000 by June?” or “Who will win the 2028 Democratic nomination?”
The contracts trade between $0 and $1. If the event happens, the contract pays $1. If it doesn’t, it pays $0. The current price therefore represents the market’s collective estimate of the probability. A contract trading at $0.65 implies a 65% chance the event occurs.
This structure is what separates prediction markets from gambling. In a traditional sports bet, the odds are set by a bookmaker. In a prediction market, the odds emerge from the aggregate behavior of all participants—buyers and sellers who each bring their own information, analysis, and sometimes their own money.
The result is a real-time probability estimate that updates continuously as new information enters the market. When Anthropic’s CEO met with the White House this week, Polymarket’s “Best AI Model” odds shifted within minutes.
How Polymarket Became the Center of Gravity
Polymarket launched in 2020 as a crypto-native prediction platform running on Polygon. It gained mainstream visibility during the 2024 US presidential election, when its odds diverged sharply from polling averages and attracted coverage from every major news outlet.
The platform uses USDC—a dollar-pegged stablecoin—for all transactions, which means every bet is denominated in dollars but settled on a blockchain. This gives Polymarket global access (anyone with a crypto wallet can participate) while sidestepping traditional banking rails.
By April 2026, Polymarket hosts markets across politics, crypto, AI, sports, and geopolitics. Its most active market—the 2028 Democratic Presidential Nominee—has over $1 billion in total volume. The “Best AI Model” market tracks whether Anthropic, OpenAI, Google, or others will hold the top spot each month, with Anthropic currently at 91.5% odds.
The platform’s appeal is straightforward: it converts dispersed information into a single number. If 10,000 people each have a piece of the puzzle, the market price synthesizes their collective knowledge into a probability that no single analyst could produce alone.
The $1 Billion Iran Betting Controversy
In April 2026, traders placed over $1 billion in bets on whether the US would invade Iran before 2027. The Financial Times reported that some bets appeared “perfectly timed”—placed just hours before geopolitical developments that shifted the odds.
The volume drew immediate scrutiny. Critics argued that betting on war is morally repugnant and that large positions could be used to manipulate public perception of conflict probability. Supporters countered that the market was simply aggregating information that governments, intelligence agencies, and defense contractors already possess.
The Iran markets also exposed a structural tension. Prediction markets function best when participants have genuine information advantages—when someone with insider knowledge of a supply chain can bet on a product launch, or when a policy expert can bet on legislation. But when that information advantage comes from classified intelligence or diplomatic cables, the ethical lines get blurry fast.
Odds on a US-Iran invasion hit 32.5% after jumping 9% in a single day, with total volume exceeding $609,000 in 24 hours. Meanwhile, the Strait of Hormuz traffic market—betting on whether oil tanker traffic would return to normal by April 30—dropped to 23.5%, an 11% decline in a day.
The Regulatory Fight: Congress vs. the CFTC
The US Commodity Futures Trading Commission has jurisdiction over prediction markets, and its stance has been inconsistent. The CFTC approved the Iowa Electronic Markets decades ago but has blocked or sued newer platforms that expand beyond academic research.
In 2026, bipartisan congressional legislation aims to create a formal licensing framework for prediction markets. The push is driven partly by the NBA’s data partnership with a prediction market operator, and partly by financial firms like Charles Schwab, which entered the prediction market space earlier this month.
The CFTC’s core objection is the Commodity Exchange Act‘s prohibition on contracts involving “gaming” or “terrorism.” Markets on whether the US will invade Iran blur the line between financial hedging and gambling on violence. The agency has asked courts to shut down several event contracts while simultaneously considering how to regulate them.
The regulatory outcome matters because prediction markets are only useful at scale. A market with 100 participants and $10,000 in volume produces noisy, unreliable odds. A market with 100,000 participants and $100 million in volume produces odds that consistently beat professional forecasters. Regulation that chokes volume kills the market’s predictive power.
Why Prediction Markets Beat Polls and Experts
Research from the University of Iowa, spanning 36 years of presidential election data, shows that prediction market prices have outperformed polls in every cycle since 1988. The mechanism is straightforward: polls measure what people say they think; markets measure what people are willing to bet on.
The distinction matters because saying something is free. Betting costs money. When someone puts $10,000 on a Trump re-election contract, they’re expressing a level of confidence that a poll respondent saying “probably Trump” never will.
Markets also aggregate faster than polls. A poll takes days to conduct and publish. A market price updates in real time. When news breaks—when Anthropic launches a new model, when the Fed hints at a rate cut, when a CEO resigns—market prices adjust within seconds.
The efficiency isn’t perfect. Markets can be manipulated by large players, they can overreact to noise, and they can persistently misprice events that are genuinely uncertain. But over thousands of markets and millions of trades, the aggregate track record is strong enough that intelligence agencies, corporate strategy teams, and hedge funds now use prediction market data as an input to their own forecasting.
The NBA, Crypto, and the Mainstream Moment
The NBA’s partnership with a prediction market operator—integrating event odds into official league data feeds—marks the first major professional sports league to formally embrace betting-adjacent forecasting tools. The deal stops short of allowing bets on individual games, focusing instead on season-long markets like championship odds and award winners.
For the crypto industry, prediction markets represent a genuine use case for blockchain infrastructure beyond speculation. Polymarket’s USDC-denominated contracts settle on Polygon, process without traditional banking intermediaries, and remain accessible to anyone with a wallet. The platform has become one of the most visible applications of crypto technology doing something that traditional finance cannot easily replicate.
Whether that mainstream moment lasts depends on regulation. If Congress creates a workable licensing framework, prediction markets could become a standard information source alongside news, polling, and expert analysis. If regulators shut them down, the forecasting function migrates offshore or underground—less transparent, less accessible, and less useful to the public.
FAQ
Are prediction markets legal?
In the US, prediction markets exist in a gray zone. The CFTC has approved some academic platforms (like Iowa’s Electronic Markets) while blocking others. New legislation introduced in 2026 aims to create a formal licensing framework. Polymarket operates from outside the US for regulatory reasons but remains accessible to US users.
How are prediction markets different from sports betting?
Sports betting odds are set by bookmakers based on their models. Prediction market prices emerge from the aggregate behavior of all participants trading contracts. The price is the probability—not a bookmaker’s estimate of it.
Can prediction markets be manipulated?
Yes, particularly in low-volume markets. A single large player can temporarily shift prices. But manipulation is self-correcting: if a market is mispriced, other traders profit by pushing the price back toward its true probability. The more liquid the market, the harder it is to manipulate.
What is Polymarket?
Polymarket is the largest crypto-native prediction market platform. It runs on the Polygon blockchain and uses USDC for all transactions. Users trade binary contracts on events ranging from elections to AI model rankings to geopolitical conflicts.