Kalshi and the Operational Limits of Prediction Markets: When the Product Becomes Regulation
During the weekend of February 28 to March 1, 2026, following the attacks by the United States and Israel in Iran that culminated in the death of Supreme Leader Ali Khamenei, prediction markets underwent a stress test that marketing cannot gloss over. In Kalshi—a federally regulated platform in the U.S.—a contract regarding whether Khamenei would be ‘out’ as a leader generated nearly $55 million in volume. In Polymarket, the equivalent surpassed $58 million. Between the two platforms, the total flow of bets tied to attacks, death, and regime changes linked to the episode reached at least $255 million, with $200 million concentrated in just four bets on Polymarket, as reported by Business Insider.
The important figure is not merely the volume; it’s the signal. When a market moves tens of millions in hours, the product ceases to be merely an interface and becomes social infrastructure: it defines incentives, distributes profits, and, ultimately, can amplify regulatory and reputational risks. In this case, the shock came not only from the geopolitical event but from how Kalshi applied its ‘death carveout’ rule to prevent the contract from becoming a direct bet on death.
Kalshi communicated that it does not offer markets resolved by death and that it does not allow markets on military attacks because such activities would be illegal under the regulatory framework for these instruments in the U.S. Nonetheless, the contract existed in a sensitive semantic zone: “out of office” is a formulation that can include resignation, dismissal, incapacitation, or death. The platform applied its ‘death carveout’: instead of paying the full value to those who had the ‘yes’ bet, it settled based on the last traded price before the confirmation of death, in addition to reimbursing fees, and announced partial refunds to those who purchased after a cutoff point communicated by its co-founder, Tarek Mansour, on X.
The reaction was swift. Users complained about unexpected losses and perceived ambiguity. Concurrently, surveillance tools like Bubblemaps flagged suspicious activity on Polymarket: new wallets supposedly earned more than $1.2 million betting on the attacks, allegedly funded in the 24 hours prior. None of this proves culpability on its own; it does, however, demonstrate an inevitable reality for any startup transforming “information” into “price”: scrutiny rises with liquidity.
The Fine Print Is No Longer Legal Protection, It’s Product Logic
Kalshi believed that the ‘death carveout’ was a brake. In practice, it functioned as a second algorithmic layer of the market: a resolution rule that changes the risk profile and expected return. From a business perspective, this is crucial. In a prediction market, the contract is not a simple agreement between parties; it is the code that defines how uncertainty is transformed into money. When the contract contains an exception that modifies the settlement at the point of greatest sensitivity, the contract ceases to be “conditions” and becomes “mechanics.”
The controversy reveals a pattern that I see often repeated in regulated startups: the temptation to push the product to the edge of what is permitted, trusting that a warning on the contract page is enough to contain the risk. That approach scales poorly. In high informational volatility scenarios—conflicts, politics, public health—users do not read like lawyers; they operate as traders. A rule that is activated only in an extreme event is perceived as a surprise clause, even if it is published.
The reported response in the headline inspiring this article—more fine print—suggests a typical defense: to strengthen the language to reduce ambiguity. It’s a comprehensible move but incomplete. The operational lesson is another: if a rule is central to the economic outcome, it must be integrated into the user's journey as an explicit product constraint, not as a paragraph. The interface should function as a risk control system: dynamic warnings, liquidation simulation under different scenarios, and trading halts when the underlying event enters a window of asymmetric information.
Here, the cost is not only reputational. In regulated markets, trust is a financial asset: it lowers customer acquisition costs, reduces disputes, and holds liquidity. When participants feel that the market “changes the rules” at the critical moment—even if not true—the liquidity evaporates or migrates. And liquidity, in this business, is the product.
Kalshi vs. Polymarket: Two Models, One Achilles’ Heel
The comparison with Polymarket is inevitable because both platforms captured massive volume in the same event. But the real contrast is not technological; it’s one of governance and regulatory perimeter. Kalshi operates as a regulated platform in the U.S. under the oversight of the CFTC, and claims that it does not allow bets on the deaths of public figures or military attacks. Polymarket, by operating offshore and without equivalent regulation, can list markets that Kalshi cannot.
This difference creates a delicate competitive incentive. The regulated platform has legitimacy but also product limits. The unregulated platform can offer “what the market wants” with less friction, capture attention and volume, and then deal with disputes through its own resolution and clarification processes. Business Insider reports that Polymarket published a clarification in a market about the “forced removal” of the Iranian leader, resolving it as “no” because the U.S. would have “assisted” in the event, and that another contract was under “final review.”
In both cases, the Achilles’ heel is the same: the resolution. In prediction markets, resolution is where public truth, sources, definitions, and timing converge. If the resolution mechanism is not perceived as predictable, the market resembles more a game of interpretable rules than a signaling instrument.
The industry already bears recent precedents of suspicions of insider trading. Business Insider notes that Kalshi suspended and fined two users for insider trading in February 2026, and in other jurisdictions, arrests have been reported tied to the use of military secrets to bet on Polymarket. Without assuming individual culpability, the strategic point is clear: when the underlying issue is geopolitical or security, the risk of asymmetric information is not a remote possibility; it is part of the terrain.
This compels an operational conclusion for startups: competitive advantage will not come just from listing more contracts but from building verifiable trust. And that requires investment in market surveillance, conservative listing policies, and a resolution design that is easy to explain, hard to manipulate, and consistent over time.
The Real Risk for a Startup: Confusing Growth with Social License
I see abundance where others see scarcity when technology reduces marginal costs and opens access. Prediction markets are exactly that: a way to convert scattered opinions into an aggregated signal with a price. But that abundance becomes fragile when the product creates socially unacceptable or politically explosive incentives.
The Iranian episode lays bare a tension that no growth committee can resolve with campaigns: there are categories of events where the market perceives a “bet” even though the contract is written as “leadership.” Kalshi attempted to design a firewall with its ‘death carveout,’ liquidating at the price before confirmation of death to avoid direct payment for death. The intention is clear: to separate “change of leader” from “death” as a liquidable event. The effect, however, was to expose that the incentive model continues to orbit around the same outcome.
From a business perspective, this impacts three fronts.
First, regulatory risk. Legislators were already pressuring the CFTC: six Democratic senators were calling for action against contracts that “incentivize physical injury or death,” according to reports. In addition, Senator Chris Murphy announced intentions to push legislation to prevent people close to power from “profiting from war.” When politics enters the table, regulatory uncertainty becomes a valuation discount.
Second, brand risk. In financial products, the brand is not an aspirational narrative; it’s the perception of fairness in execution. A screenshot of a user feeling “robbed” circulates faster than any legal explanation. Reputation is especially delicate when the underlying issue is a real tragedy.
Third, design risk. The market does not forgive ambiguities. If a contract allows multiple interpretations, arbitration shifts from “information” to “definitions.” This attracts participants not seeking prediction but social litigation or dispute resolution. And that dynamic degrades the price signal, which is the value proposition.
The strategic exit is not to prohibit everything; it’s to recognize that the product needs guardrails. Technology can democratize access to the signal, but governance must prevent market efficiency from becoming operational waywardness.
Towards Predictive Markets That Scale with Control: Less Ambiguity, More Design
Kalshi responded by applying rules and promising more contractual clarity. It’s a first step, but the deeper direction for the sector is more demanding: to convert governance into competitive advantage.
A regulated market that wants to grow without becoming a crisis magnet needs to develop three capabilities.
The first is predictable resolution. The definition of the event, valid sources, and the logic of liquidation must be consistent and tested with extreme scenarios before listing the contract. The death exception, for instance, cannot be a detail; it must be shown as an explicit bifurcation of economic outcomes.
The second is surveillance as a product, not just as compliance. If the market is exposed to events where insider information exists by nature, the platform must elevate the investment in detecting anomalous patterns and act with clear procedures, as Kalshi has already done in other reported cases. In practice, this translates into operational cost, but also into an asset: it reduces the risk of regulatory captures and retains sophisticated liquidity.
The third is a discerning listing policy. The industry has learned that listing more markets does not equate to creating more value. Polymarket had 187 markets related to Iran, many with little volume. Proliferation can be a mirage of breadth: it increases the dispute surface area, fragments liquidity, and multiplies risks.
From my lens, this sector is undergoing the transition where the digital ceases to be a promise and becomes a structure. Prediction has digitalized, scaled rapidly, and now faces its phase of institutional disruption. The winner will not be whoever writes more fine print, but whoever transforms clarity, surveillance, and design limits into a market experience that empowers human judgment without rewarding asymmetry.
The current phase is one of disappointment and entry into disruption: the market discovers that the marginal cost of listing contracts tends to zero, but governance and trust still carry a real cost, and technology must empower the human element through clear rules that democratize the signal without degrading responsibility.
