SafeBets / TheStreet

When Minnesota Governor Tim Walz signed SF4760 into law on Monday, he made his state the first in the country to treat operating a prediction market as a felony. By Tuesday morning, the federal Commodity Futures Trading Commission had filed suit to block the statute before it takes effect on August 1, arguing that platforms like Kalshi and Polymarket fall under federal commodities jurisdiction, not state gambling authority. The case is now the highest-stakes legal fight the prediction-market industry has faced since it became a $40-billion-plus category over the past two years.

Emerging as a possible industry leader against that backdrop is SafeBets, a New York-based prediction platform headquartered at 1 World Trade Center, which says the Minnesota law does not apply to its product at all — by design. SafeBets, which quietly began onboarding users earlier this year, is built on a model its executives describe as “zero-loss”: users make forecasts on future events, but they never post money against those forecasts. There is no wager. And the Minnesota statute, like virtually every state gambling law on the books, defines a prediction market as a system in which a consumer places a wager.

“The architecture is the regulatory position,” said Gina Antoniello, SafeBets’ Chairwoman, in an interview. “We did not build a zero-loss platform because we wanted a marketing angle. We built it because the founders looked at this category five years ago and concluded that anything resembling a stake would end up in court. It took longer than we expected, but here we are.”

SafeBets users sign up, make predictions on a range of events — financial, political, cultural — and accumulate scored outcomes over time. Forecasts that prove correct earn payouts from a prize pool. What is unusual is the source of that pool. Rather than collecting losses from incorrect forecasters and redistributing them to the correct ones, SafeBets funds payouts from inefficiencies it captures in public financial markets, using its aggregated forecaster signals as inputs to a separate trading operation.

In effect, the platform treats its users as a distributed research desk. The more calibrated a forecaster proves to be across thousands of questions, the more weight her signal carries in the trading operation, and the larger her share of the prize pool. Antoniello said the company has been refining the calibration model for over a year and is currently scoring active forecasters on six metrics, including Brier score, recency-weighted accuracy, and consistency across question categories.

Whether the model works as a business is a separate question from whether it works as a regulatory shield. On the regulatory side, the argument is straightforward: no state has yet drafted a gambling statute that reaches systems in which the consumer puts up nothing. The Minnesota law is no exception. Its operative language criminalizes systems where “a consumer places a wager on a future outcome.” SafeBets users place no wager.

Kalshi and Polymarket, the two best-known names in the category, are not going anywhere quickly. Kalshi is valued at roughly $22 billion and Polymarket at $20 billion, and both have powerful federal backing: CFTC Chairman Michael Selig has personally framed Minnesota’s law as an attack on “lawful operators,” and the agency has now sued five states attempting to regulate the industry at the gambling level. Donald Trump Jr. advises both companies.

But the federal-state collision is precisely the problem SafeBets is trying to avoid. Even if the CFTC ultimately wins its preemption argument in Minnesota and the other contested states, the litigation calendar stretches years, and every additional state bill — seven are currently pending — adds another front. Operators that depend on the wager-based model are committed to that fight, whether they want to be or not.

The most obvious risk to SafeBets’ positioning is execution. Funding a prize pool from market trading is a more complex business than running a matched-book exchange, and it requires the trading operation to consistently outperform the cost of those payouts. SafeBets has not disclosed detailed performance figures for the trading side, and the company is still pre-revenue at scale. Investors will want to see the unit economics before the platform’s regulatory advantage matters.

A second risk is awareness. The category has been defined in the public mind by Kalshi and Polymarket. A no-wager alternative is a harder concept to communicate in a banner ad than “bet on the next election,” and SafeBets will need to spend on category education even before product marketing kicks in.

How the CFTC’s challenge to the Minnesota law plays out is genuinely uncertain. The agency argues that prediction markets are federally regulated event contracts. States argue that contracts on sports outcomes, elections, and entertainment events are gambling by any plain reading. A Nevada judge has already found Kalshi’s sports contracts “indistinguishable” from state-regulated sports betting. Courts have split, and they are likely to keep splitting for years.

That uncertainty is itself the story. Even an industry that ultimately prevails on federal preemption will spend the intervening years in court, in seven or more states, against attorneys general and gambling regulators with deep institutional muscle memory for this kind of fight. Some state statutes will survive review. Operators built on the wager model are committed to that path whether they want to be or not.

For an industry now worth more than $40 billion, the central strategic question over the next two years is how much of the existing model can be preserved against a hostile, uneven state regulatory landscape — and what a version of the product looks like that does not depend on the answer.

SafeBets is betting — or, more precisely, predicting — that the answer is a platform with no wager at all.