Introduction

On July 29, 2003, two U.S. senators walked into a press conference and ended a Pentagon research program in a single news cycle. The program, called the Policy Analysis Market, would have let a small pool of registered traders buy and sell contracts on political and economic events in the Middle East. Within about 24 hours of the press conference, the Defense Department had abandoned it.

Twenty-three years later, the mechanism the Pentagon could not defend operates at commercial scale. Polymarket, the largest prediction market platform in the world, reported roughly $26.2 billion in trading volume in the first quarter of 2026 alone. It holds a federal license. The parent company of the New York Stock Exchange has committed up to $2 billion to it and distributes its data to institutional clients.

The lineage here is intellectual, not corporate. No DARPA money flowed to Polymarket, and no personnel connect the two. What connects them is a single idea — that market prices are an efficient way to aggregate scattered information about uncertain future events — and a set of objections to that idea that Congress raised forcefully in 2003 and has not resolved since.

Two questions follow. Why did the objections stop a government program but not a private one? And what happens to the data these platforms generate about the people trading on them?

Background: What DARPA Actually Built

The Policy Analysis Market grew out of a DARPA research effort called FutureMAP, housed in the Information Awareness Office. That office was directed by retired Admiral John Poindexter, who was already a lightning rod for his role in the Iran-Contra affair and for a separate surveillance program, Total Information Awareness.

DARPA contracted with Net Exchange, a small San Diego firm, to build the exchange. Economist Robin Hanson of George Mason University, one of the intellectual architects of prediction markets, was involved in the design. In December 2002, DARPA solicited related research proposals; by the summer of 2003, research teams at several institutions had received awards of roughly $100,000 each.

The design was narrower than its critics described. Traders were to be a limited pool — reporting varies between about 100 initial testers and up to 1,000 registered participants — trading contracts on economic and political indicators such as regional GDP, military activity, and government stability in specific Middle Eastern states. Registration was scheduled to open in early August 2003.

In simple terms: the theory holds that a person who is willing to risk money on a forecast reveals more about what they actually believe than a person filling out a survey. Aggregate enough of those bets and the resulting price behaves like a probability estimate.

The 48 Hours

Senators Ron Wyden of Oregon and Byron Dorgan of North Dakota went public on July 28 and 29, 2003, describing the program in terms that made it politically indefensible. Wyden called the concept of a federal betting operation on atrocities “ridiculous,” and Dorgan characterized it as “useless, offensive and unbelievably stupid.”

On the Senate floor, Majority Leader Tom Daschle raised the objection that has proven the most durable: traders in ordinary markets try to influence their own investments. He asked how long it would take before someone traded in a way designed to bring about the outcome they had bet on.

The coverage collapsed almost instantly. Fewer than a half-dozen mostly neutral articles had mentioned the program in the preceding months; roughly fifty largely negative pieces appeared on July 29 alone. DARPA’s press office was reportedly caught flat-footed. Wyden secured an amendment to the defense spending bill requiring congressional approval before the program could proceed. Poindexter resigned in August.

The aftermath produced an irony that economists still cite. An offshore exchange, Tradesports, listed a contract on whether Poindexter would be out by the end of August. Researchers Justin Wolfers and Eric Zitzewitz documented that the contract’s price spiked sharply within minutes of insider reports circulating — hours before the resignation became public knowledge. The market worked. That was precisely the problem.

What Happened Next: The Government Built It Anyway

The part of this history that rarely surfaces in current coverage is that the intelligence community did not abandon the concept. It moved it behind a classification barrier.

The Intelligence Advanced Research Projects Activity, the intelligence community’s counterpart to DARPA, ran the Intelligence Community Prediction Market on classified networks beginning around 2010. Participants were cleared government employees and contractors drawn from across the intelligence agencies. They traded non-monetary points on questions about geopolitical outcomes, and the resulting price served as a consensus forecast. The program was reportedly discontinued in 2020.

IARPA also funded a series of public forecasting efforts — the Aggregative Contingent Estimation program, the ForeST program and its SciCast market, and the Hybrid Forecasting Competition. The Good Judgment Project, which emerged from ACE under researchers Philip Tetlock and Barbara Mellers, produced forecasts that outperformed intelligence analysts with access to classified material.

The CIA’s own academic journal, Studies in Intelligence, published a favorable assessment of prediction markets as an analytic tool in 2006, three years after PAM’s cancellation.

The record therefore does not show that Congress rejected the method. It shows that Congress rejected a version of the method that used real money, involved outside traders, and could be described in a headline as betting on terrorism.

Policy Explanation: How the Regulatory Position Reversed

Polymarket was founded in 2020 by Shayne Coplan and built on Polygon, a blockchain network. Its early trajectory ran directly into federal derivatives law.

In 2022, the Commodity Futures Trading Commission settled an enforcement action against the company, then operating as Blockratize Inc., over unregistered event-based binary options. The order imposed a $1.4 million civil penalty and required the company to wind down noncompliant markets. U.S. users were blocked.

The reversal came in stages. In July 2025, reporting indicated that both the Justice Department and the CFTC closed post-election investigations into the company without charges. Polymarket then acquired QCX, an entity already holding a CFTC license as a designated contract market — the same regulatory category as established futures exchanges. On November 25, 2025, the CFTC issued an amended order of designation permitting Polymarket to operate an intermediated trading platform under the full set of requirements applicable to federally regulated U.S. exchanges. A U.S.-facing product followed, with access to American users expanding through early 2026.

The practical result is two platforms sharing one brand. The global exchange runs on-chain, with wallet-based onboarding and no identity verification required by default. The U.S. exchange runs through registered intermediaries and requires full identity verification before a user can deposit funds.

That structural split is where the accountability questions concentrate.

Examples: The Objections Arrive on Schedule

Daschle’s 2003 objection — that people with advance knowledge would trade on it — stopped being hypothetical in January 2026.

On January 3, three newly created Polymarket wallets took positions on whether Nicolás Maduro would be removed from power in Venezuela by the end of the month. One wallet, identified on-chain as 0x31a5, staked roughly $34,000 when the market priced the outcome at about 6 percent. Hours later, U.S. special forces captured Maduro in Caracas. The on-chain analytics firm Lookonchain calculated the three wallets’ combined profit at $630,484, with the largest single position clearing more than $400,000. The accounts had no meaningful trading history outside Venezuela-related contracts.

No enforcement action established insider access, and alternative explanations circulated — including one trader’s claim that he had inferred an imminent operation from unusual late-night activity at restaurants near the Pentagon. But the pattern recurred. Roughly twelve new accounts reportedly took positions ahead of the first airstrike on Iran, netting about $330,000.

The first confirmed case came from outside the United States. In February 2026, Israeli authorities indicted an IDF reservist and a civilian on serious security charges, alleging they used classified operational information to place Polymarket bets on the timing of Israeli strikes against Iran.

Scale is not trivial. A Bloomberg Businessweek review of activity on Polymarket’s global exchange found roughly 34,000 transactions flagged as potential insider trades by the analytics platform Polysights between August 2025 and June 2026.

The Surveillance Question: Who Watches, and Who Buys

Here the accountability picture diverges sharply from the way these platforms are usually described.

Polymarket’s global exchange is frequently characterized as private because it does not require identification. The opposite is closer to true. Every trade, position, and transfer is recorded permanently on a public blockchain and tied to a wallet address. Anyone — a competitor, a journalist, an employer, a foreign intelligence service — can reconstruct a wallet’s complete trading history at no cost and with no legal process. What the platform does not collect is a name to attach to it. Everything else is permanently public.

In simple terms: the absence of KYC does not make a trader anonymous. It makes them pseudonymous, and pseudonyms are frequently resolvable through deposit patterns, timing, and links to identified accounts elsewhere.

On April 30, 2026, Polymarket announced a partnership with Chainalysis, the blockchain analytics firm, to deploy on-chain monitoring and investigative software against insider trading and manipulation. The companies stated the tools would produce blockchain-verified evidence for use in law enforcement and regulatory inquiries. The announcement was framed as market integrity infrastructure, and it is. It is also a surveillance layer built on a dataset that was already fully exposed.

The commercial question is who buys the resulting signal. On October 7, 2025, Intercontinental Exchange announced a strategic investment of up to $2 billion in Polymarket at roughly an $8 billion pre-investment valuation, completing the commitment with a further $600 million injection in March 2026. Under the agreement, ICE became the exclusive global distributor of Polymarket’s event-driven data to institutional investors. In February 2026, ICE launched a Polymarket signals and sentiment product delivering normalized probability feeds through the same consolidated infrastructure that carries NYSE pricing data.

The aggregated market signal, in other words, is now a licensed institutional data product. The underlying trader-level activity remains freely readable by anyone with a block explorer.

The KYC Gap

The identity question is where regulatory pressure has concentrated. The U.S. platform requires name, date of birth, residential address, Social Security number, and government-issued identification, with that data shared among the CFTC-regulated exchange and approved futures commission merchants as law requires. The global platform requires none of it by default.

Polymarket tightened that posture in May 2026, blocking VPN access more aggressively and pressing high-volume traders toward verification, with unverified accounts facing restrictions. A company engineering executive publicly stated that broader identity requirements applied to a beta product rather than the core platform. Kalshi, the CFTC-regulated competitor, has taken a different approach, screening candidates out of their own election markets and athletes out of their own sports markets, and collecting employment information from users in certain markets.

On May 22, 2026, House Oversight Committee Chair James Comer opened an investigation into both Polymarket and Kalshi, writing to Coplan and Kalshi’s Tarek Mansour to demand records on identity verification practices, geographic restriction enforcement, and how the platforms flag anomalous trading. In July 2026, reporting indicated the CFTC had opened a separate inquiry into Polymarket’s marketing and consumer protection practices — a matter distinct from the legality of the trading itself.

Legislation has followed the same track. Representative Ritchie Torres introduced the Public Integrity in Financial Prediction Markets Act of 2026 on January 9, drawing more than thirty Democratic cosponsors including former Speaker Nancy Pelosi. The bill would bar federal elected officials, political appointees, executive branch employees, and congressional staff from trading event contracts tied to government policy or political outcomes when they hold or could obtain material nonpublic information through their duties. In May, Torres and Representative Seth Moulton introduced a companion measure covering campaign staff, consultants, and pollsters. Moulton had already imposed an office-level ban on his own staff in March.

Impact

Three consequences are already measurable.

Institutional adoption is outpacing rulemaking. Prediction market volumes have grown sharply since early 2024, and a Bernstein analysis in April 2026 projected the sector could approach roughly $1 trillion in volume by 2030. Kalshi raised approximately $1 billion at a $22 billion valuation in early 2026. Proprietary trading firms have begun deploying automated agents into these markets.

The national security exposure runs in both directions. A Council on Foreign Relations analysis published in April 2026 noted that more than $1 billion has traded on geopolitical prediction markets since 2022, and that foreign intelligence services can read and potentially manipulate that public order flow as easily as domestic analysts can. Meanwhile, a June 2026 article in the U.S. Naval Institute’s Proceedingsargued the opposite case — that joint intelligence doctrine should be updated to treat prediction markets as a legitimate open-source intelligence input.

Enforcement authority remains unsettled. Traditional insider trading law is built around securities and a duty owed to shareholders. Event contracts on government action do not fit that structure cleanly, which is why the response has taken the form of new legislation rather than existing enforcement.

Analysis: Which 2003 Objections Survived

Sorting the original objections against the current record produces an uneven result.

The moral objection — that a federal agency should not operate a betting venue on assassinations — was the one that killed the program, and it applied specifically to government sponsorship. It has no purchase on a private exchange. Nothing in federal law prevented a startup from building the same mechanism.

The manipulation objection, that traders would act to bring about their own positions, remains largely theoretical at scale. No documented case has established a trader causing a geopolitical event to profit from a contract.

The information leakage objection has been substantiated. This was Daschle’s concern, and the Venezuela trades, the Iran positions, and the Israeli indictments all describe the same failure mode: individuals with access to operational information converting it into money, in public, in minutes.

One objection has emerged that no one raised in 2003, because the technology did not exist. PAM would have been a closed government system with vetted participants and classified oversight. Polymarket’s global exchange is a permanent, public, machine-readable ledger of who bet what and when. The program Congress rejected as a surveillance risk would have been substantially more private than the market that replaced it.

Conclusion

The most accurate framing is not that Polymarket descends from DARPA. It is that both are implementations of the same economic proposition, separated by two decades and by the question of who bears the political cost when it goes wrong.

In 2003, a Pentagon research office bore that cost immediately and completely. In 2026, the cost is diffuse — spread across a federally licensed exchange, a Fortune 500 data distributor, a blockchain analytics contractor, two congressional investigations, and a set of bills that have not passed.

The mechanism the government could not defend in public is now infrastructure. The oversight question is no longer whether to permit prediction markets. It is who holds the trading records, who can read them, who can buy the derived signal, and whether anyone with a security clearance can be stopped from trading on what they know.

Key Takeaways

DARPA’s Policy Analysis Market was announced and effectively canceled within roughly 24 hours in July 2003 after Senators Ron Wyden and Byron Dorgan publicly condemned it.
The intelligence community continued the research privately. IARPA ran the Intelligence Community Prediction Market on classified networks from approximately 2010 until 2020, alongside publicly funded forecasting programs.
There is no institutional link between DARPA’s program and Polymarket. The connection is a shared economic premise, not funding, personnel, or technology.
The CFTC fined Polymarket $1.4 million in 2022 and issued it an amended order of designation as a federally regulated exchange in November 2025 after the company acquired a licensed derivatives venue.
Suspiciously timed trades preceding the January 2026 capture of Nicolás Maduro produced $630,484 in profit across three new wallets. Israeli prosecutors filed the first confirmed insider trading indictments tied to a prediction market in February 2026.
The global platform’s lack of identity verification does not create privacy. Every position is permanently public on-chain; only the name is missing.
ICE holds exclusive global rights to distribute Polymarket’s event-driven data to institutional investors, delivered through the same feed infrastructure as NYSE pricing data.
House Oversight opened an investigation into Polymarket and Kalshi in May 2026. Federal legislation restricting insider trading in event contracts has been introduced but not enacted.

Sources

Hanson, Robin. “The Policy Analysis Market: A Thwarted Experiment in the Use of Prediction Markets for Public Policy.”Innovations, 2007. direct.mit.edu
Wolfers, Justin and Eric Zitzewitz. “Prediction Markets.”Journal of Economic Perspectives, 2004. csc2.ncsu.edu
Congressional Record, Senate floor statements on the Policy Analysis Market, July 29, 2003. sgp.fas.org
NBC News. “Pentagon kills ‘terror futures market,'” July 29, 2003. nbcnews.com
Central Intelligence Agency, Studies in Intelligence. “Using Prediction Markets to Enhance US Intelligence Capabilities.”cia.gov
Federal Foresight Community of Interest. “Three IARPA Forecasting Efforts: ICPM, HFC and the Geopolitical Forecasting Challenge,” January 2018. ffcoi.org
Intercontinental Exchange. “ICE Announces Strategic Investment in Polymarket,” October 7, 2025. ir.theice.com
Polymarket. “Polymarket Receives CFTC Approval of Amended Order of Designation,” November 25, 2025. prnewswire.com
CoinDesk. “Polymarket taps Chainalysis to bring Wall Street-level oversight to crypto prediction markets,” April 30, 2026. coindesk.com
CoinDesk. “Congress Probes Polymarket and Kalshi Over Fears Government Employees Are Trading on Secret Info,” May 22, 2026. coindesk.com
PBS NewsHour. “A $400,000 payout after Maduro’s capture put prediction markets in the spotlight,” January 2026. pbs.org
Front Office Sports. “New Bill Would Bar Insider Trading in Prediction Markets,” January 9, 2026. frontofficesports.com
Office of Rep. Ritchie Torres. “Reps. Torres And Moulton Introduce Bill To Ban Campaign Staff From Insider Trading On Political Prediction Markets,” May 2026. ritchietorres.house.gov
Irregular Warfare Initiative. “Precision-Guided Predictions: The Intelligence Risk in Prediction Markets,” March 2026. irregularwarfare.org
Council on Foreign Relations. “The Ultimate Price of Prediction Markets,” April 2026. cfr.org
U.S. Naval Institute Proceedings. “Prediction Markets Could Be a Valuable OSINT Tool,” June 2026. usni.org
Bloomberg Businessweek. “Polymarket and Kalshi Grapple With a New Era of Insider Traders,” 2026. bloomberg.com
PYMNTS. “CFTC Investigation of Polymarket Broadens Compliance Questions for Prediction Markets,” July 8, 2026. pymnts.com