How Does Kalshi Make Money? The Spread, The Fees, And The 5% You Do Not See
How does Kalshi make money? From a published fee schedule you can read in about a minute, and that is the only line Kalshi itself banks. But trading there costs you twice, because there is a second charge you cannot read anywhere: a spread baked into the prices themselves, collected by the traders on the other side of your order. You pay both, so both belong on one page. The fee half takes a minute to read and is still widely misdescribed, because the fee is a curve, not a flat rate. The price half is the one nobody had done the work on: as far as we can find, no one has published a measured figure for the spread built into Kalshi's prices, so we measured it ourselves. Across 770 daily weather events, one six-rung ladder each, priced mid-life with our own recorders, the built-in charge came out to a median of 5.0% per ladder. This page walks both halves, and by the end you will know where the five cents hides, who collects it, and why the arithmetic that exposes it looks better on paper than it can ever be in practice.
The Quick Answer
Kalshi charges a published trading fee of 0.07 × price × (1 − price) per contract when your order crosses the spread and fills instantly, and a quarter of that rate when your order rests and waits to be filled. The cost that never appears as an itemized fee: on the daily weather ladders we track, the six ask prices sum to a median of 1.0500 across 770 events priced mid-life, a measured 5.0% carried in the prices rather than listed anywhere as a charge. The fee curve most explainers get backwards, the full measurement, and the six-rung arithmetic that makes naive backtests look brilliant are all below.
The Half You Can Read: Kalshi's Fee Schedule
Kalshi's trading fee is a formula, not a percentage of your stake: 0.07 × price × (1 − price) per contract for takers, the orders that cross the spread and fill instantly, and 0.0175 × price × (1 − price) for makers, the resting orders that wait for someone else to cross. The gap between the two rates is 4x for the same market. On a 45-cent contract it is the difference between 1.73 cents and about 0.43 cents, which is why the order type you choose matters more than most new traders realize. So does Kalshi take a cut of your winnings? Not as a commission on profit. The charge lands per contract when you trade, and it scales with the price of the contract in a way almost nobody expects.
That shape is the most useful thing to know about the schedule. Kalshi's fee peaks at 50 cents and falls to zero at both ends of the price range, because price × (1 − price) is largest on a coin-flip and collapses as a contract approaches certainty. A 95-cent near-certainty and a 5-cent long shot cost the same to trade, and each costs just under a fifth of what a 50-cent contract does.
| Contract Price | Taker fee per contract | Maker fee per contract |
|---|---|---|
| 5C | 0.33c | 0.08c |
| 50C | 1.75c | 0.44c |
| 95C | 0.33c | 0.08c |
Formula values at the peak and the two ends. The full line-by-line schedule, including how orders are actually billed, lives in our breakdown of the fee schedule.
The row worth sitting with is the middle one. Even the most expensive fill on the board, a taker order on a dead coin-flip, costs 1.75 cents per contract, and resting that same order cuts it to under half a cent. If the schedule were the whole cost of trading on Kalshi, the visible price of a trade would look almost trivial. It is not the whole cost, and the rest of it does not appear on any statement.
The Half You Cannot: A Measured 5% Inside The Prices
Here is the audit that the fee schedule invites. A Kalshi daily temperature market is a ladder of six bands for one city and one day, each band a contract that pays $1 if the day's temperature lands in it. (If event contracts are new to you, start with how prediction markets work.) Exactly one band settles yes. Not usually one, but always: in 1,793 daily weather events we tracked, exactly one rung won 100.0% of the time. Six rungs, one winner, which means the average rung carries a 16.7% chance before anyone looks at a forecast.
That structure turns the ladder into a probability distribution you can check. If the ask prices were pure probabilities, the six of them would sum to exactly $1.00: one certain winner, priced in full. They do not. Across 770 events priced mid-life, the six ask prices summed to a median of 1.0500, and the result reproduced independently on 744 live order-book snapshots. A quarter of ladders summed below 1.020. A quarter summed above 1.076.
| Six-Rung Ask Sum | Measured |
|---|---|
| 25Th Percentile | 1.020 |
| Median | 1.050 |
| 75Th Percentile | 1.076 |
Sample: 770 events priced mid-life, reproduced independently on 744 live order-book snapshots.
Bookmakers have a word for the amount above 100%: the overround, the same thing sportsbook bettors call the vig when it is built into a two-sided line. (How an exchange's version differs from a sportsbook's is its own subject, covered in an exchange is not a sportsbook.) In plain terms, when you buy a 30-cent rung at the ask, you are paying a shade more than the ladder's own arithmetic says the outcome is worth, and the surcharge across all six rungs runs about a nickel per dollar of payout, which is roughly 4.8% of the money you actually put up. That 5% is the house edge, and it is in the prices rather than on your statement.
One precision matters here: the five percent is not a second Kalshi fee. Kalshi the company banks the fee schedule above; the nickel goes to whoever is resting on the other side of your order, the makers standing on the ladder all day, as payment for being there the moment you want to trade now. Every venue with a bid and an ask works this way. What is unusual is not that the spread exists but that, as far as we can find, nobody had measured it here before. From your side of the screen, though, the accounting is cold comfort: both charges come out of your price.
A Worked Example: Selling A Whole Ladder
The 5% is not an estimate that depends on a forecast model: you can check it with arithmetic alone. Sell all six rungs of a ladder and exactly one will settle against you, so your return is the sum of the six prices minus the winner's $1 payout, identically, whatever the thermometer does. One premise before the numbers, because it decides which side of the book you are on: selling at those prices means resting your own offers at the ask and waiting for buyers to cross to you. That is why the maker fee rate applies below, and it is also the catch in advance. Hit the six bids for instant fills instead and you are on the paying side of the same spread, with the identity running against you.
On 1,566 complete ladders we measured, a separate and larger sample than the 770 events above, selling all six rungs returned +6.03 cents per ladder, and that figure is gross, before fees. We know that without checking a fee statement, because the identity's residual came out to +0.0000 cents: a clean zero is only possible when no fee has been taken out yet, so the +6.03c figure certifies itself as before-fees. The net belongs beside it, always: maker fees across six rungs run 1.09 cents at Kalshi's published 0.0175 rate, so the net is +4.94 cents per ladder, and quoting the gross alone overstates the result by 22%. The access constraint belongs beside it too: all six rungs are simultaneously fillable only 12.5% of the time, which in our recorder data worked out to roughly 6.5 fillable baskets a day, capping the whole identity at about $1,822 a year at 20x size. That dollar figure is a ceiling on how much of the six cents exists to be touched, not a projection of anyone's results. The arithmetic is ironclad; the access to it is not. Six cents is the arithmetic; the 12.5% is whether you ever get to touch it.
Why This Page Is Not Telling You To Sell Ladders
Read that paragraph again and notice what it is: an identity, not a strategy, and this page is not telling you to sell ladders. The identity is precisely why a naive backtest of ladder-selling looks brilliant; the price archives contain the sums, the sums sit reliably above a dollar, and a spreadsheet will happily multiply the six cents by 365. What the archive cannot contain is whether all six of your orders would have filled together, and 87.5% of the time they would not. Liquidity decides whether you can trade at all, and no quote archive records the fills you never got.
There is also the standing shape of risk on the selling side, and it never changes: selling an unlikely outcome collects a small premium and risks most of a dollar, so roughly speaking, one loss erases the premiums from about 15 wins. That asymmetry, not any hit rate, is what makes position sizing the whole game. We say all of this as participants rather than spectators: Stokastic trades these markets and holds positions in them, we do not publish performance figures, and our own live log is short, currently negative, and far too small to confirm or refute anything. The current picture stays on our standing hub covering how these markets work, which is rebuilt through the day precisely so this page never has to freeze a number that will go stale.
SO Where Does Your Money Actually Go?
Back out to the whole picture, because the two halves now fit together, and only one of them bends. Trade a Kalshi weather ladder as a taker and you pay twice: once to Kalshi, through a published formula that tops out at 1.75 cents on a coin-flip rung, and once to the traders resting on the other side, through the median 5.0% (770 events) built into the six ask prices. Of those two charges, only one is negotiable, and not the one people expect. You cannot argue with the spread, but you can decline to cross it: rest your order instead of taking, and the visible fee falls by 4x while the invisible one flips from a toll you pay into the queue you are standing in. That single choice, repeated, is most of what the fee schedule rewards, and the same make-versus-take decision also runs through our guide to trading NFL markets on Kalshi. None of this is a scandal; a spread is what every market charges for immediacy, and every exchange and sportsbook has its version. Kalshi itself operates with broad, state-specific availability under federal oversight. The difference now is that its spread is measured, off the same six-rung structure (1,793 events, one winner every time) that makes each rung a 16.7% base rate, and knowing the size of the toll before you pay it is most of what separates a trader from a customer.
The same toll exists on every sportsbook line you have ever bet, and there it has a name: the vig. The difference is that you can shop it, because the books quote against each other all day, and comparing the same bet across 20+ sportsbooks to see which one charges you least is exactly the job OddsShopper does. Our free expert picks publish the reasoning next to every selection, free, every day. When you want the full price-comparison toolkit, OddsShopper Pro comes with a free week trial, and the code HOUSEEDGE20 takes 20% off your first month if you stay past the week.
Disclosure and fine print. Stokastic trades event markets on Kalshi and holds positions in them; the measurements on this page come from our own recorders. We have no affiliate or commercial relationship with Kalshi, though we do carry sign-up offers for some other prediction-market and betting platforms. Kalshi weather contracts are CFTC-regulated event derivatives traded on a designated contract market, not sportsbook wagers, and they can lose their full value. 18+, available where Kalshi operates, with broad, state-specific availability under federal oversight; the risk of loss is real and, on the side we trade, individually large. This page is part of an open research log of a strategy we have not yet proven, and nothing here is trading advice.
