The end of Polymarket's free era: a multi-party battle over "fees"

Until now, only crypto and sports contracts were charged. After this expansion, the only surviving “free zone” is left with geopolitics.

By Kabuda, Deep Tide TechFlow

In the 2024 U.S. presidential election, Polymarket caught fire on zero fees and $3.3 billion in presidential election betting volume, becoming synonymous with prediction markets worldwide.

A year and a half later, the company decided to start charging everyone.

Starting March 30, Polymarket officially expanded Taker fees to nearly all contract categories—politics, finance, economics, culture, weather, technology, with not a single category left out. Previously, only crypto and sports contracts were subject to charges. After this expansion, the only remaining “free zone” is geopolitics.

Free lunch ends here.

I. How are the rates collected?

Polymarket didn’t choose the traditional fixed-commission model. Instead, it adopted a “dynamic probability pricing” mechanism: fees fluctuate with a contract’s win probability. The closer the probability is to 50% (when the market is least certain), the higher the fee. When the outcome is close to certain (probability trending toward 0% or 100%), the fee approaches zero.

Specifically, the peak fee rates by category are:

• Crypto contracts: 1.80% (up from 1.56% previously)
• Economics: 1.50%
• Culture/Weather: 1.25%
• Politics/Finance/Technology: 1.00%
• Sports: 0.75% (up from 0.44% previously)
• Geopolitics: 0% (the only free category)

Here’s an example: for a $50 sports contract, if the probability is exactly 50/50, the fee rises from the original $0.22 to $0.38. The increase is even larger for crypto contracts, directly squeezing the real returns of high-frequency traders.

At the same time, Polymarket rolled out a Maker rebate program. Not all fees are platform profit; instead, they’re returned daily to liquidity providers (market makers) in the form of USDC. The rebate percentages differ by category: up to 50% for finance, and 25% for sports. The logic is clear: charge retail traders (Takers), subsidize market makers (Makers), and use fees to drive the liquidity flywheel.

II. Why start charging now?

The answer is hidden in three numbers.

First, Polymarket’s trading volume over the past 30 days is about $9.55 billion. Based on an estimated blended effective fee rate under the new pricing, the platform’s daily revenue will reach between $0.8 million and $1.0 million—about $300 million annualized. For a company that doesn’t yet have a stable revenue model, this money is the foundation for survival.

Second, ICE, the parent company of the New York Stock Exchange, has just completed a commitment to invest a total of approximately $2.0 billion in Polymarket. The first $1.0 billion will be in October 2025, and in March 2026 there will be an additional $0.6 billion in cash plus an acquisition of $40 million worth of old shares. When that deal was signed, Polymarket’s valuation was about $8.0 billion; now, according to reports, the platform is preparing a new round of financing at a valuation close to $20.0 billion. When investors put in that much money, they want to see revenue.

Third, Polymarket’s competitor Kalshi is already charging. Its annualized revenue has reached $1.5 billion, and its valuation has soared to $22.0 billion as well. If Polymarket doesn’t charge anymore, it’s essentially dressing up the competitor with its own free traffic—educating users for free, only for those users to turn around and trade on a paid platform with deeper liquidity.

There’s also a background that can’t be ignored: Polymarket has just signed an exclusive multi-year partnership agreement with MLB, reportedly worth as much as $300 million. It had already reached partnerships with the NHL, MLS, and UFC before. Signing with a professional sports league means the platform must operate commercially. You can’t shake hands with a major league and tell investors, “We haven’t figured out how to make money yet.”

III. The “crackdown” from Congress

Polymarket chose a delicate timing to start charging.

One week before the fee expansion, California Democratic Senator Adam Schiff and Utah Republican Senator John Curtis teamed up to introduce the “Prediction Markets Are Gambling Act,” which would bar any prediction contracts related to sports events from being listed on trading platforms that are registered by the CFTC.

Schiff put it plainly: “Sports prediction contracts are sports gambling—just under a different name.”

Curtis’s concern is more specific: Utah’s constitution bans all gambling, but Polymarket and Kalshi’s prediction contracts are still available in all 50 states, bypassing all state-level regulation.

This isn’t an isolated case. In the same week, Oregon Democratic Senator Jeff Merkley introduced a more aggressive “STOP Corrupt Bets Act.” It wouldn’t just ban sports; it would also ban election-related, government-action-related, and military-action-related prediction contracts. The Arizona attorney general has already filed criminal lawsuits against Kalshi, alleging unlicensed operation of gambling businesses. Meanwhile, Nevada courts issued temporary restraining orders against Kalshi, and have issued similar rulings against Polymarket as well.

Multiple pieces of legislation moving forward at the same time—prediction-market regulation is facing the most intense regulatory offensive since the industry’s inception.

But the market isn’t too nervous right now. On Polymarket, contracts about whether the “2026 act banning sports prediction markets” will pass show a probability of only 9.5%. These bills still have to clear many hurdles—committee hearings, votes in both chambers, and presidential signature—so given the crowded agenda in the current Congress, the odds of actually landing are very low.

IV. The shadow of manipulation hasn’t lifted

Beyond the fee controversy, Polymarket has an even more troublesome problem: accusations of market manipulation and insider trading have never stopped.

In January of this year, a newly created account made pinpoint bets before Venezuela’s President Maduro was arrested, netting more than $400,000. In March, before the coordinated attacks by the U.S. and Israel on Iran, the on-chain analytics firm Bubblemaps found six newly created wallets profited by $1.2 million within a few hours through related contracts. Israeli authorities even arrested two people accused of using confidential military intelligence to place bets on prediction markets.

Earlier than that, a market with over $7.0 million in trading volume—“Whether Ukraine would agree to Trump’s minerals deal”—was forcibly ruled as “Yes” without any official confirmation, sparking widespread user protests. One user even directly called it a “Polyscam” on social media.

A user named Folke Hermansen posted detailed accounts of multiple manipulation cases on X. The core accusation is: Polymarket’s rulings depend on the UMA token voting mechanism, and two whales control more than half of the voting power—one address holds 7.5 million out of 20 million UMA tokens. Regular users can’t really challenge the ruling outcomes.

Polymarket’s response was to introduce a “strengthened version of market integrity rules,” explicitly banning trading using stolen confidential information, building positions using insider knowledge, and participating in trades by parties who can influence outcomes. At the same time, the platform announced a collaboration with Palantir and TWG AI to build a market monitoring system.

Are these measures enough? In an interview on CNBC, Senator Schiff answered: “Saying ‘this is our policy’ isn’t enough. The key is whether you have the means to truly implement it.”

V. The $20 billion betting game

When you connect all the clues, Polymarket’s situation looks like a classic startup equation:

ICE pours in nearly $2.0 billion—it wants prediction-market data as a new kind of financial infrastructure. ICE CEO Sprecher said plainly in the earnings call that this isn’t venture capital behavior; ICE’s return logic is to integrate prediction-market data into its own workflows, driving data sales revenue. MLB grants an exclusive partnership—it wants prediction markets to help drive user growth for the league. Congress wants regulatory power and control over the gambling industry. Users want zero fees and fair rulings.

Among these needs, at least two sets are fundamentally contradictory.

The zero-fee era attracts users and traffic, but it can’t sustain a company valued at $20 billion. Charging can bring in revenue, but it may push price-sensitive retail traders to competitors. DraftKings has already announced it will build its own prediction-market market-making division, while FanDuel has teamed up with CME Group to enter the space. In 2026, the two major traditional gambling giants are expected to invest a total of several hundred million dollars into prediction markets. And if Congress truly passes a sports ban, Polymarket’s fastest-growing category right now will be cut off directly.

Polymarket made a smart hedge: keep geopolitics contracts free. This both sticks to the platform’s positioning as a “public prediction tool” and signals to Congress that it’s not just a gambling platform—it provides valuable collective intelligence.

But the line between “valuable collective intelligence” and “gambling” has never been determined by technical architecture. It’s determined by political bargaining.

Polymarket’s own contracts provide the answer: the chance the ban passes is below 10%. But then again, if you trust Polymarket’s contracts to determine Polymarket’s fate, that in itself is an interesting recursive problem.