On February 28, 2026, the United States and Israel launched coordinated strikes against Iran. Within minutes, prediction markets lit up.

Polymarket processed over half a billion dollars in Iran-related contracts in a single trading session. Kalshi saw $54 million flow into a single market on whether Ayatollah Ali Khamenei would be removed as Supreme Leader. Across both platforms, traders were pricing regime change, escalation timelines, nuclear risk, and downstream economic effects — all before most cable news anchors had finished their opening monologues.

For a brief window, prediction markets did exactly what their defenders have always claimed they could do: they became the fastest, most granular source of real-time probability on a rapidly evolving geopolitical crisis.

Then everything broke.

The Speed Was Real

Before we get to the failures, the speed deserves acknowledgment — because it was extraordinary.

Traditional intelligence analysis operates on cycles. Analyst teams gather information, synthesize assessments, debate internally, and publish conclusions. Even the fastest desks at major banks measure their response time in hours. Government agencies measure theirs in days.

Prediction markets repriced the probability of Khamenei's removal in under sixty seconds after the first reports of strikes. Downstream markets — oil price thresholds, Fed rate expectations, S&P 500 correction probability — began moving within minutes, as traders on Kalshi and Polymarket independently processed the implications of the same event for different economic outcomes.

This is the core value proposition of prediction markets as an intelligence tool. Not that they're always right, but that they synthesize distributed information faster than any centralized system can. When thousands of traders with different information sets, different analytical frameworks, and different risk tolerances all trade simultaneously, the resulting price is a probability estimate that updates in real time.

During the Iran crisis, that estimate moved faster than wire services, faster than financial terminals, faster than official government communications. For anyone watching multiple platforms simultaneously, the signal was extraordinary.

The problem is that almost nobody was watching multiple platforms simultaneously. And the people who were there got very different pictures depending on where they looked.

The Fragmentation Was Brutal

Here is what the Iran week actually looked like for a trader or analyst trying to use prediction markets as an intelligence source.

Polymarket had the deepest liquidity on geopolitical outcomes — regime change, strike timing, escalation probability. Its international exchange operated outside CFTC jurisdiction, which meant markets could exist that regulated platforms couldn't touch. Traders wagered on nuclear detonation timelines. They bet on assassination outcomes. The breadth was unmatched.

Kalshi had the strongest infrastructure for downstream economic effects — recession probability, Fed rate decisions, S&P 500 thresholds. As a CFTC-regulated exchange, it couldn't list the same raw geopolitical markets Polymarket could, but it offered something Polymarket didn't: a direct bridge between event outcomes and financial market expectations.

The problem was that these two pools of intelligence existed in complete isolation from each other.

A trader on Polymarket could see that the probability of Khamenei's removal had spiked to 78%, but couldn't see what Kalshi traders were pricing for downstream economic effects. A trader on Kalshi could see recession odds climbing, but couldn't see the geopolitical trigger data on Polymarket that was driving the move.

The signal wasn't on any single platform. It was in the space between them.

And that space — the synthesis layer — didn't exist.

The problem was that these two pools of intelligence existed in complete isolation from each other.

Table 1: Cross-Platform Intelligence Map: What Each Venue Could — and Couldn't — Tell You During the Iran Crisis

The synthesis problem in one line: A trader who needed both geopolitical trigger data (Polymarket's strength) and economic downstream pricing (Kalshi's strength) had no way to see them together. The complete intelligence picture required both platforms — and no tool existed to bridge them.

The Resolution Disaster

Speed and fragmentation were the appetizer. The resolution crisis was the main course.

When Khamenei was confirmed dead in the strikes, the two largest prediction market platforms handled the outcome in opposite ways.

Polymarket resolved its "Khamenei out as Supreme Leader" market as Yes. The man was dead. He was, by any reasonable interpretation, no longer Supreme Leader. Traders who had bet Yes collected their winnings.

Kalshi invoked a "death carveout" — a clause buried in the market's fine print stipulating that outcomes directly tied to death would not be honored as standard resolutions. Instead of paying out Yes at full value, Kalshi settled positions based on the last traded price before Khamenei's death was confirmed and refunded all trading fees. Traders who had correctly predicted the outcome received partial payments. Many received nothing close to what they expected.

The backlash was immediate and fierce. Traders accused Kalshi of changing the rules mid-game. A class action lawsuit was filed. The company's CEO posted a lengthy explanation on X, acknowledging that the death carveout "was not prominent enough" in the market's presentation.

This wasn't an isolated incident. Just weeks earlier, both platforms had faced a nearly identical crisis over the Cardi B Super Bowl performance market — $57 million wagered on the same question, opposite answers across Kalshi and Polymarket about whether her appearance constituted a "performance."

Two events. Two resolution disasters. Combined volume exceeding $100 million in disputed outcomes. And a pattern that reveals something fundamental about where prediction markets are in their evolution.

Table 2 —  The Resolution Disaster 

Resolution Scorecard: When the Same Event Produces Different Answers

What the pattern reveals: Resolution isn't a technical problem. It's a structural one. Each platform writes its own rules, interprets its own definitions, and resolves its own markets — with no shared standard, no cross-platform consistency, and no way for traders to know in advance how the same event will be treated across venues.

The Integrity Gap

The Iran week also exposed what may be the industry's most dangerous vulnerability: the near-total absence of market integrity infrastructure.

A Polymarket account trading under the username "Magamyman" made over $553,000 on bets placed just before the strikes began. Blockchain analytics firm Bubblemaps identified six suspected insider wallets that collectively earned $1.2 million on suspiciously timed Iran wagers. The trades were placed hours — in some cases, minutes — before bombs hit Tehran.

In traditional financial markets, this kind of trading pattern would trigger immediate investigation. Surveillance systems would flag the anomaly. Compliance teams would freeze accounts. Regulators would subpoena records.

In prediction markets, almost none of that infrastructure exists.

Polymarket's international exchange operates outside CFTC jurisdiction. Users trade through anonymous cryptocurrency wallets. There is no systematic surveillance for unusual trading patterns. There are no market-wide circuit breakers. There is no cross-platform visibility that would allow regulators — or anyone else — to see coordinated suspicious activity across venues.

Kalshi, as a regulated exchange, has somewhat stronger safeguards. The company disclosed that it has opened roughly 200 insider trading investigations and escalated about a dozen into active cases. But its surveillance is limited to its own platform. It cannot see what's happening on Polymarket, or on any of the dozens of smaller prediction market venues that have emerged in the past year.

Congress noticed. Senator Chris Murphy announced legislation to ban insider trades related to the Iran conflict. Representatives Blake Moore and Salud Carbajal introduced a bill that would block prediction markets from offering contracts on war, sports, and other categories entirely. The End Prediction Market Corruption Act would bar the president, members of Congress, and their families from trading event contracts.

The regulatory response may be heavy-handed. It may miss the point. But the trigger was real: prediction markets are growing faster than the integrity infrastructure that institutions require before they'll participate at scale.

What the Stress Test Actually Revealed

Every technology stack has a moment when real-world stress exposes the gap between what exists and what needs to be built. For prediction markets, Iran was that moment.

Table 3: Iran Stress Test: Where Prediction Markets Passed and Failed

The pattern is clear. Prediction markets passed the speed test — the one thing they were designed to do. They failed on every infrastructure layer that institutions require before they'll participate at scale. And the layer that could address fragmentation, resolution risk, and integrity — the intelligence layer — doesn't exist yet.

The Layer Nobody Built

In our previous analysis of the prediction market stack, we mapped five infrastructure layers: Exchanges, Distribution, Clearing, Resolution, and Intelligence. The Iran week tested every layer simultaneously — and the intelligence layer, which barely exists, is where the most critical failures concentrated.

Not because the data wasn't there. It was. Across platforms, prediction markets generated an extraordinarily rich, real-time dataset on geopolitical probability, economic downstream effects, and market sentiment.

The failure was in synthesis. No single platform could give a trader or analyst or institution a complete picture of what prediction markets were actually saying about Iran. To get that picture, you needed to manually monitor multiple venues, reconcile different contract structures, understand different resolution rules, assess different liquidity profiles, and somehow process all of that fast enough to act before the information was already priced in.

That's not a user interface problem. That's an infrastructure gap.

The intelligence layer — the layer that sits above the exchanges and turns fragmented, cross-venue data into synthesized signal — is what was missing during the $1 billion Iran week. It's what was missing during the Cardi B resolution debacle. It's what's missing every time a trader opens four browser tabs and tries to manually reconcile what different platforms are saying about the same event.

Bloomberg built this layer for traditional financial markets and generates over $10 billion in annual revenue — more than most of the exchanges whose data it aggregates. The prediction market equivalent doesn't exist yet.

That's the gap. That's what needs to be built. And the Iran stress test just proved why it matters.

What Comes Next

The prediction market industry is at an inflection point. The Iran week compressed a decade's worth of structural questions into seven days.

The regulatory response is coming. Whether it's targeted legislation on war-related contracts or broader CFTC rulemaking — which is reportedly planned for 2026 — the rules of the game are about to change. Platforms that have operated in regulatory gray zones will face new constraints.

The institutional interest is real but conditional. Hedge funds are already ingesting prediction market data as an alternative dataset. The partnerships between Polymarket and Dow Jones, between Kalshi and CNN, signal that mainstream financial infrastructure wants access to this signal. But institutional capital demands reliability, consistency, and transparency that the current fragmented landscape can't provide.

The intelligence layer is becoming inevitable. As the number of venues increases, as resolution rules diverge, as the volume of cross-platform signal grows, the need for a synthesis layer that aggregates, normalizes, and scores prediction market data across venues becomes more acute — not less.

Iran didn't break prediction markets. It revealed where the infrastructure is strong, where it's fragile, and where it doesn't exist yet.

The speed is the proof of concept. The fragmentation is the opportunity. The resolution and integrity gaps are the problems. The intelligence layer is the solution.

This is the sixth installment in the Assymetrix Intelligence Brief series, examining the structural evolution of prediction markets.

Previous: "The Prediction Market Stack: Who Owns Each Layer and Why It Matters"

Assymetrix is building the intelligence and synthesis layer for prediction markets — cross-platform aggregation, signal extraction, and the data infrastructure this industry is missing.