The world's largest prediction market, Polymarket, is starting to charge fees! Behind it is a calm game of regulation, survival, and timing.

1. It suddenly started charging money, but you might not have noticed at all

You might have seen pages like these:

• “Probability that Trump wins the 2024 election: 51.3%”
• “Probability that the Fed cuts rates in March: 68.7%”
• “LPL Spring Finals, odds that BLG wins: 1.39”

This isn’t a gambling website, and it isn’t media commentary either. It’s a special kind of presence in the Web3 world: a prediction market.

Put simply, it’s a “vote” mechanism backed by real money. If you believe something will happen, you buy a “Yes” contract. If you believe it won’t happen, you buy a “No” contract. The price moves in real time, and the final number reflects a “collective judgment” formed by thousands of people putting their money behind it.

And Polymarket is currently the hottest and most active on-chain prediction platform globally, with the data being cited the most. By providing a clean web interface, it lets users trade directly with the USDC stablecoin.

On January 6, 2026, it quietly updated its website. In its documentation, it added a page called “Trading Fees,” and announced that starting immediately, markets of the “15-minute crypto asset up/down” type would charge a fee, capped at 3%.

When the news broke, many long-time users’ first reaction was: “Huh? Weren’t they always free before? Then how did they make money back then?”

That question hits a truth in the Web3 world that’s often overlooked: a cool-looking technical product, to truly survive, has never depended on code and ideals alone.

2. It went viral on hot topics, but its survival is decided by regulation

Polymarket really has taken off multiple times:

• In the 2022 Qatar World Cup, users bet on “Argentina to win,” and contract prices surged nonstop.
• In the 2023 LPL Spring Finals, esports fans traded team outcomes on the platform in real time.
• In the 2024 U.S. election, the peak daily trading volume topped $2.7 billion. Even The New York Times cited it as a source.

But what truly determines whether it can keep operating has never been those flashy events. It comes down to two words: regulation.

After it was founded in 2020, Polymarket quickly gained support from well-known venture capital firms like Founders Fund, which is backed by Peter Thiel. At one point, it even planned to roll out nationwide across the United States. But in January 2022, a single enforcement action from the U.S. Commodity Futures Trading Commission (CFTC) shut it down:

The binary contracts it offered—things like “Who will win Real Madrid vs Barcelona” or “Whether the Fed will cut rates”—fall under regulated swap transactions, which must obtain licenses for a “designated contract market” (DCM) or a “swap execution facility” (SEF)—and it didn’t.

So what happened? Polymarket agreed to pay a $1.4 million penalty and shut down all compliance-risk markets facing U.S. users. On the surface it looked like it exited. In reality, it was a strategic contraction: moving the entity out of the U.S., turning funding channels into on-chain settlement, while still keeping services open to the rest of the world—including U.S. users.

What’s interesting is that exiting the U.S. actually made it more “mainstream.”

During the 2024 election, it became an “unofficial dashboard” that people worldwide used to track shifts in public sentiment. Before writing articles, media would check it. Traders would reference it for modeling. Researchers would also route their analysis of public sentiment through its API.

And the real turning point came in November 2025, when the CFTC formally approved its DCM application. This means—no longer is it “an innovation project skating the edge,” but it has obtained a formal “license badge” within the U.S. financial regulatory system.

This move to charge fees wasn’t a spur-of-the-moment idea. It was the first step after getting that badge.

3. It was free for six years—not because it wasn’t making money, but because it was waiting for the “right time to earn with peace of mind”

You might not know this: most prediction markets have already been charging fees for a long time—common rates range from 0.5% to 3%. But since Polymarket launched in 2020, it has charged zero fees for all users and all markets.

That sparked lots of speculation: Is it staying alive on venture capital? Is it selling data? Are big players behind the scenes covering the costs?

The real answer is more practical: it was betting on a time window.

The value of a prediction market isn’t how much profit you make on any single trade. Instead, it depends on whether enough people participate, and whether participation is frequent enough, to produce真实, stable, and credible price signals. And “zero fees” is the most direct and effective way to attract users.

Over six years, it succeeded in three things:

• In high-attention events like politics, sports, and crypto, it became the de facto “default pricing center.”
• Its price data was repeatedly referenced by Bloomberg terminals, academic papers, and hedge fund strategies, becoming a de facto standard.
• It accumulated a complete probability dataset spanning years, across cycles, events, and regions—an economic moat that no new platform can buy by spending money.

In other words, it traded away the money it could have collected for something more valuable: liquidity, influence, and data assets.

And the decision to start charging on January 6, 2026 is the natural outcome of this long-term strategy:

• Only for markets of the “15-minute crypto up/down” type—high-frequency, short-term, and easily disrupted by bots.
• Dynamic fee rates: the closer the price is to 50% (the harder it is to judge), the higher the fee; the closer it is to 0% or 100% (the more certain), the lower the fee—even down to zero.
• All fees don’t go into the platform’s pocket. Instead, every day they are returned in full in USDC to the market makers (the people providing buy/sell quotes).
• The goal is straightforward: incentivize more order placement, narrow the bid-ask spread, and enable faster fills even during sharp crashes or explosive rallies.

Some say it’s to crack down on high-frequency “wash trading” bots. Others think it’s to filter out fake trades. Still others point out that, at its core, it’s a stress test: within the boundaries allowed by regulation, verifying whether a fee model can improve market quality without harming user experience.

It didn’t become “commercial.” It simply finally got to “do business seriously.”

4. Small scope, big upside; just starting out, already under pressure

Don’t underestimate the significance of this fee limited to a single category.

Based on data compiled by the on-chain data analysis firm Gate Research on the Dune platform:

• In the first two weeks after fee launch, Polymarket had accumulated about $2.19 million in trading fees.
• At the current pace, weekly revenue is about $730k, and a static annualized estimate reaches $38 million.

This is only for the “15-minute crypto up/down” niche. Polymarket’s current coverage includes:

• U.S. and global political elections
• Top sports events like the World Cup, NBA, and LPL
• Macro events like Fed rate decisions and CPI releases
• Long-cycle topics such as cryptocurrencies, real estate, and AI technology progress

The profit space is far from fully opened. But the other side of the coin is this: compliance is never something you achieve once and for all.

Getting a CFTC DCM license only means it passed the “federal-level exam.” But the U.S. is a federal system, and states have the right to set their own financial and gambling regulations. In mid-January 2026, Tennessee’s sports betting regulator issued a stop order to Polymarket and similar platforms like Kalshi, clearly requiring:

“Immediately stop providing sports event contracts to residents of this state, or face civil penalties and even criminal charges.”

Similar challenges exist around the world:

• In Japan, the Financial Services Agency (FSA) explicitly lists event contracts as prohibited business.
• In the UK, the FCA requires licensing + high deposits + strict anti–money laundering reviews.
• In mainland China, all prediction markets are inaccessible, and policy also explicitly prohibits them.

So Polymarket’s next step isn’t a sprinting expansion—it’s ongoing adaptation:

• Establish localized compliance entities in different jurisdictions.
• Clearly define product design boundaries between “financial instruments” and “entertainment activities.”
• Explore partnerships with traditional financial institutions, turning probability data into inputs for risk control models.

Can it become a “evergreen tree” in the Web3 world? The answer isn’t whether the technology is more advanced, but whether it can find a sustainable middle path among regulation, users, and business.

Prediction markets give us a rare perspective: when the world is full of uncertainty, at least we can know this—right now, how many people worldwide are willing to put real money behind the idea that “this will happen.”

This consensus might not be correct, but it is real enough. And Polymarket’s fee move isn’t the end of the story—it’s the beginning of it truly growing up as a real service.