Share
On May 29, 2026, the Commodity Futures Trading Commission (CFTC) approved KalshiEX, LLC to list the BTCPERP Contract, a perpetual futures product referencing the spot price of Bitcoin. The approval made Kalshi the first company in American history to offer regulated perpetual futures.
CFTC Chairman Michael Selig, who was confirmed in December 2025 following his appointment by President Trump, called it “a major step forward in delivering on President Trump’s goal of cementing America as the crypto capital of the world,” as reported by Fortune on May 29.
The development is the sharpest possible proof of the argument this piece makes. Kalshi did not classify its event contracts as futures when it added BTCPERP. It created a separate product line with its own regulatory approval, mechanics, and market function. The company that pioneered prediction markets in the US has implicitly acknowledged, in the most formal regulatory language available, that what it was building before and what it launched on May 29 are different instruments. The market just did not have clean language for why.
This piece supplies it. The central argument is falsifiable: if event contracts, sports bets, and financial futures share the same pricing mechanism, participant incentive structure, and information function, then treating prediction markets as a distinct asset class produces no analytical or regulatory benefit. The evidence shows they do not share any of the three.
A financial futures contract obligates parties to transact in an underlying asset at a predetermined price on a specific future date. The structural elements that define this instrument are the underlying, the delivery mechanism, and the continuous price path between initiation and settlement. CME crude oil futures obligate delivery of 1,000 barrels of WTI.
Treasury futures reference notional Treasury securities. Even cash-settled contracts reference a continuously observable price series whose evolution matters throughout the contract’s life.
Perpetual futures are a modification of this structure, removing expiry while retaining the continuous underlying reference. The BTCPERP contract tracks Bitcoin’s spot price with no termination date.
A funding rate, charged every eight hours per Kalshi’s launch announcement, anchors the contract price to the underlying spot market.
The instrument Kalshi built functions as continuous risk transfer against a live Bitcoin price, which is exactly what offshore platforms like Binance and Hyperliquid have been offering at scale.
Reuters placed perpetual futures volume at $61.7 trillion in 2025, up 29% from 2024. Coinbase CEO Brian Armstrong noted on May 29 that US traders had been “locked out of roughly 80% of global crypto markets” that consist of perpetuals and options, per his X post that day.
Big day for our US-based traders, and for Coinbase.
Until now, US users have been locked out of ~80% of global crypto markets (perpetual futures and options). But not anymore!
Coinbase is the first and only regulated platform able to connect US users to global crypto options… pic.twitter.com/7EdDUN3Yn6
— Brian Armstrong (@brian_armstrong) May 29, 2026
None of these characteristics apply to event contracts.
A Kalshi contract on “Will the Fed cut rates in September 2026?” has no underlying asset to deliver or hedge. There is no continuous price path because the event is discrete and non-recurring. The contract settles at $1 or $0 on a single future date and then ceases to exist. There is no funding rate, no maintenance margin against a live spot price, no mechanism to roll positions forward.
The CFTC’s decision to require a separate product approval for BTCPERP under Regulation 40.3, distinct from Kalshi’s existing DCM authorization for event contracts, is the regulatory embodiment of this structural difference. One approval covers the prediction market. A different approval covers the futures product. The regulator drew the line.
The distinction between prediction markets and sports betting is structurally more fundamental than the regulatory debate acknowledges.
A sportsbook is a house-banked operator. It sets prices, acts as counterparty on every trade, and builds its margin directly into the odds. On a standard -110/-110 NFL spread, the implied probabilities sum to approximately 104.76%, meaning the 4.76% is the book’s guaranteed margin on every dollar wagered, per Fox Sports’ vig analysis published in April 2026.
On player props and niche markets, sportsbook margins often reach 8–10%, while four-leg same-game parlays can carry effective holds above 20%. The goal is not perfect prediction but balanced action that guarantees the vig.
As a result, sportsbook odds do not reflect true probabilities. They reflect the bookmaker’s estimate adjusted for margin and risk management. A -135 favorite implies a 57.4% chance after vig, while the underlying market probability is typically lower once that margin is removed.
Prediction markets charge an explicit 0-2% trading fee with no embedded margin, per Gambling Insider’s structural comparison published in January 2026. A $0.62 Kalshi contract represents the market’s collective estimate of the probability that an event will occur. No margin has been embedded to guarantee the platform’s profits regardless of the outcome. The platform profits from the fee, not from being the counterparty. This is the peer-to-peer exchange model that sportsbooks have never used, and regulatory frameworks built for sportsbooks are not designed to evaluate it.
Prediction markets price non-recurring binary events through peer-to-peer epistemic aggregation. Traders do not bet against the house. They do not transfer exposure to a continuously observable underlying layer. They express probability estimates about discrete future outcomes and trade against counterparties with different estimates. The market price becomes a calibrated aggregation of dispersed private information.
This produces a distinct information output. Combined monthly volume on Kalshi and Polymarket reached approximately $24 billion in April 2026, per Pew Research Center’s May 2026 analysis, surpassing the roughly $14 billion average monthly handle of US legal sportsbooks in 2025. Kalshi’s January 2025 monthly volume of $175 million grew to $24.98 billion by May 2026, per Token Terminal data, more than a 100-fold increase in twelve months.
May 2026 became the sector’s second-largest month on record at $25.7 billion on April 7, 2026, a nearly 13-fold increase from the $2 billion recorded in March 2025.
The Iowa Electronic Markets, which have operated since 1988, showed that prediction markets were closer to the actual election outcome than polls 74% of the time, outperforming polls in every election when forecasting more than 100 days in advance.
An Arxiv paper published in 2026 found that Polymarket’s forecasts outperformed those of three major poll aggregators, including Silver Bulletin and The Economist, in predicting the 2024 presidential result, particularly in swing states. The mechanism producing this accuracy, financial incentives converting private information into tradeable probability estimates, has no structural equivalent in futures or sportsbooks.
The strongest challenge to the “distinct asset class” case comes from the record of accuracy. If prediction markets systematically misprice probabilities, their information aggregation function is compromised, and the case for treating them as a separate epistemic instrument weakens.
The CEPR study by Bürgi, Deng, and Whelan, published in 2026 and analyzing over 300,000 Kalshi contracts, is the most rigorous dataset available. Its findings require a precise statement: Kalshi’s prices are broadly calibrated, meaning a 50-cent contract wins approximately 50% of the time, and accuracy improves as the resolution date approaches. But they display a systematic favorite-longshot bias.
A 10-cent contract wins materially less often than 10% of the time. The average pre-fee return across all contracts is 20.0%, with takers losing approximately 32% and makers approximately 10%.
This is a limitation, but it does not make prediction markets the same as sportsbooks. In sportsbooks, favorite-longshot bias comes from bookmaker pricing. In prediction markets, it stems from trader behavior and the tendency to overvalue low-probability outcomes, as documented by Kahneman and Tversky.
The mechanism is structurally distinct. And the fact that sophisticated makers earn substantially better returns than unsophisticated takers is the microstructure of a financial exchange, not the uniform house-edge extraction of a casino.
Kalshi’s CEO, Tarek Mansour, framed the BTCPERP launch with explicit language in the company’s announcement: “Prediction markets were the first chapter. Perpetual futures are the next chapter.” He described event contracts as “a photograph of what the world thinks right now” and perpetuals as “a film, continuously updated, never ending, always present.”
That metaphor is analytically precise. An event contract captures a collective probability estimate at a specific moment about a specific discrete outcome. A perpetual tracks a continuous underlying with no resolving event. The information function, the participant incentive, and the pricing mechanism are all different.
The CFTC institutionalized this distinction on May 29 by approving BTCPERP via a separate regulatory pathway, distinct from Kalshi’s existing event contract authorization.
Alongside Kalshi’s approval, the CFTC also issued a no-action letter to Coinbase Financial Markets covering certain perpetual futures products to be routed through Coinbase Bermuda. Both Polymarket and Bitnomial have announced their own perp moves. The instrument category is being formally established within the US regulatory architecture as a separate category from both event contracts and traditional futures.
State gaming commissions argue that Kalshi’s event contracts are gambling, applying a framework designed for house-banked operators to a peer-to-peer exchange. Meanwhile, the CFTC applies futures-market logic to event contracts, imposing delivery and underlying-asset concepts that binary event contracts inherently lack.
The $24+ billion monthly prediction market that exists today, operating at 13 times the volume of twelve months prior, is generating a price signal about future discrete events that neither futures markets nor sportsbooks are designed to produce. That is not a regulatory classification problem. It is a new instrument, and it needed a new name before Kalshi gave it the second one.
Share
preferred on
