From Subsidies to a Fee Economy: How Polymarket's On-Chain Flows Changed in Six Months

How Polymarket moved from a subsidized marketplace to a fee-generating one in six months, and what public settlement data reveals about where those fees go. An analysis by PolyScalping, reconstructed directly from on-chain records, January–July 2026.

In its first six months of charging fees, Polymarket generated $131.42M in observable trading fees. Of that, $72.63M (55%) can be traced on-chain to reward distributions back to users; the remaining $58.79M (45%) is not identifiable as direct trader rewards in on-chain data. The redistribution rate is declining month over month. These figures reflect on-chain flows only, not internal accounting, and not all reward distributions are necessarily fee-funded.

Key findings

• $131.42M in trading fees has been recorded on-chain since fees launched in January 2026, in just over six months.
• $72.63M (55%) is traceable on-chain as reward distributions to users, across four streams. The remaining $58.79M (45%) is not identifiable as direct trader rewards in on-chain data, where it ultimately goes is not observable on-chain.
• Cumulative fees surpassed cumulative historical reward distributions in April 2026, total fees recorded matched every reward the platform had distributed across roughly two prior years of subsidy.
• The redistribution rate is declining: cumulative rewards were 69% of fees on May 1, 60% by June 1, and 56% by July 1.
• Fee volume grew +206% in a single month (April 2026) and reached a record $2.07M in a single day (July 7, 2026).
• 1.24M wallets have paid fees, but concentration is extreme: the median payer has paid $0.99, the average is $106, and the top 100 wallets account for 24% of all fees.
• Reward distributions total $74.69M across five streams (including a holding-yield program funded by deposit interest, not fees), paid to 649K wallets, only about half the number that have paid fees.

All figures are reconstructed from public on-chain records, not Polymarket's internal ledgers. See methodology and limitations below.

From subsidy era to fee economy

For most of its existence, Polymarket ran on subsidy. Between 2023 and the end of 2025, it distributed rewards, LP incentives, rebates, referrals, while charging traders nothing. That changed in January 2026, when trading fees went live.

The transition was faster than almost anyone expected. By April 4, 2026, barely three months after fees launched, cumulative fee revenue surpassed cumulative historical reward distributions. Two years of reward spend, matched by a single quarter of fees.

April's +206% jump is the fastest-growing month on record. The single biggest fee day came on July 7, 2026, when traders paid $2.07M in fees in 24 hours.

The redistribution rate is declining

The headline 55% is a snapshot, and it is trending down. Measured cumulatively, the ratio of reward distributions to fees has fallen every month since fees scaled:

In two months, the redistribution rate fell from roughly 69% to 56%. Fees are growing faster than reward distributions, and the gap is widening.

Read in Polymarket's favor, the context is important: for roughly two years before fees launched, the platform distributed rewards while collecting nothing, so current fees are partly offsetting prior subsidy rather than representing margin. But the direction of travel is clear, a shrinking share of fees is matched by observable on-chain rewards as volume grows.

Where the fees go on-chain

Polymarket's own interface does not break this down. Reconstructed from on-chain data, the $131.42M splits as follows:

Traceable to reward distributions, $72.63M (55%), across four streams:

• Maker rebates: $29.23M (22.2%)
• LP rewards: $27.20M (20.7%)
• Referrals: $10.90M (8.3%)
• Taker rebates: $5.29M (4.0%)

Not identifiable as direct trader rewards on-chain, $58.79M (45%).

This 45% figure needs care, and this analysis does not overstate it. It is fees minus observable reward distributions, not profit. It excludes operating costs, infrastructure, development, compliance, market-maker agreements, and any value that never settles on-chain, some of which may benefit users without appearing as a reward. Where the 45% ultimately goes is not visible on-chain and is not something this analysis claims to measure.

Two further caveats keep the accounting honest. First, the $2.06M holding-yield program is excluded from this split, it is funded by deposit interest, not trading fees. Second, not every reward stream should be interpreted the same way. Maker rebates, taker rebates, and referral rewards are directly tied to trading fees. LP rewards may draw on Polymarket's own capital or other incentive budgets rather than being mechanically funded by the same fee pool. The $72.63M is therefore best read as observable reward distributions, not as a one-to-one recycling of fees.

Who actually receives rewards

On the payout side, Polymarket has distributed $74.69M across five reward streams, spread over 15.4M individual transactions to 648,912 recipient wallets.

The two sides don't match in reach: 1.24M wallets have paid fees, but only 649K have ever received a reward, a little over half. The reward side, like the fee side, is narrower than the raw participation count suggests.

A market of whales and dust

The headline count of 1.24M fee-paying wallets suggests broad participation. The distribution says otherwise.

• Median fee paid: $0.99, half of all payers have paid less than a dollar.
• Average fee paid: $106, the gap between median and average is the signature of extreme concentration.
• Top payer alone: $3.42M, 2.6% of all fees, from a single wallet.
• Top 100 wallets: $31.01M, 24% of all fees, ever.

The five largest fee payers to date:

A small cohort of high-volume traders drives a disproportionate share of fee volume. No individual small account moves the needle; the concentration at the top is steep.

Why this matters

For traders, the takeaway is concrete: fees are now a first-order cost, and rebate eligibility has real dollar value. For a high-volume account, the difference between paying full taker fees and qualifying for maker rebates or LP rewards can run into five or six figures a year, enough that fee optimization is now a core part of trading strategy, not a detail.

The bottom line

Polymarket can no longer be understood purely as a subsidized marketplace. It now runs a substantial fee stream, tens of millions monthly at peak, funded disproportionately by a few hundred high-volume traders, with a declining share matched by observable on-chain rewards.

But the more significant shift is not the size of the fee pool. It is that a previously opaque marketplace economy can now be analyzed through public settlement data, fee generation, redistribution, and concentration read directly from the chain rather than inferred. That is the real story: not one headline percentage, but the fact that Polymarket's flows have become measurable at all.

Methodology & data verification

Every figure is reconstructed from public on-chain records on Polygon, not from Polymarket's internal ledgers. The reconstruction covers January 2026 (fee launch) through July 8, 2026.

What is measured. Taker fees are identified from fee transfers emitted when trades settle against Polymarket's exchange contracts. Reward distributions are aggregated end-to-end from on-chain Polygon Transfer events across all five streams, LP rewards, maker rebates, taker rebates, referrals, and holding yield, traced from known reward-distribution addresses to recipient wallets. No third-party aggregator is used and no data is scraped from Polymarket's interface.

Known limitations.

• "Not identifiable as direct trader rewards on-chain" is fees minus observable reward distributions. It is not net profit, it excludes operating costs, infrastructure, staff, legal, market-maker agreements, and any off-chain flows, some of which may benefit users.
• The reconstruction captures value that moves on-chain. Incentives settled off-chain, netted internally, or paid outside the identified contracts would not be visible.
• Reward-stream attribution depends on correctly identifying distribution contracts. Addresses are validated against known Polymarket reward mechanisms, but the classification is a reconstruction, not an official disclosure, and not all distributions are necessarily fee-funded.
• Cumulative figures update continuously as new blocks are decoded, so point-in-time totals may differ slightly from the live dashboards.

Figures can be cross-referenced against the live, continuously-updated breakdowns:

• Trading Fees: https://polyscalping.org/stats/fees
• Fees vs Rewards: https://polyscalping.org/stats/fees-rewards
• Fees Leaderboard: https://polyscalping.org/leaderboard/fees

Data current as of July 8, 2026. Figures update continuously as new fees are decoded on-chain.

This analysis was produced by PolyScalping, an on-chain analytics platform for Polymarket traders. Media and researchers are welcome to cite these figures with attribution to PolyScalping.