Prediction markets are starting to look less like one-off bets on events and more like tools for pricing the macro picture. Kalshi’s latest week showed both at the same time: a filing to list 58 U.S. stock perpetual futures, and a fast repricing in contracts tied to what the Federal Reserve does next.
The report, written by Alea Research and translated by TechFlow, said Kalshi submitted the filing to the U.S. Commodity Futures Trading Commission on Sept. 18 and self-certified the relevant rule sections. Over that same nine-day stretch, open interest on the exchange climbed from 1.221 billion contracts to 1.417 billion, a 16.0% gain, though roughly half of that increase came from football parlay markets that reset every Sunday.
Kalshi seeks approval for 58 stock perpetual futures
On Sept. 18, Kalshi filed with the CFTC for approval to list 58 U.S. stock perpetual futures. The set includes 55 common stocks, Alibaba ADR, and the SPY and QQQ funds. The names in the filing run from Apple to Exxon Mobil, and they also include SpaceX.
Each contract covers 100 shares. Minimum customer margin is fixed at 15.50% of current market value, and the exchange, not the trader, sets that requirement. The report said these products sit in the security futures category, which puts them under both Securities and Exchange Commission and CFTC supervision. Kalshi has registered by notice as a national securities exchange to trade them.
During the period covered by the report, none of the contracts had begun trading, no launch date had been set, and Kalshi had not made a public announcement about the filing.
Ninth Circuit ruling put sports contracts back in focus
Two days before the stock perpetual filing, the Ninth Circuit said in Blue Lake Rancheria v. Kalshi that “The sports event contracts offered on Kalshi’s contract market are, in substance, sports gambling, whether or not Kalshi calls them swaps.” The court reversed the denial of a preliminary injunction and remanded the case.
But the report was clear about what the ruling did not do. No product was ordered to stop trading. Nothing was taken down. That matters because sports and football parlays make up 63.8% of Kalshi’s open interest.
Traders paid $3.50 billion over eight days, with Sunday the busiest session
Across the eight days where cash and contract records overlapped, traders paid $3.50 billion, up from $3.34 billion in the prior eight-day period, a 4.6% increase. Reported contract volume on the exchange rose 12.3% over that same span.
The busiest day for cash activity was Sunday, Sept. 13, when turnover hit $547.3 million, or 15.6% of the eight-day total. That session lined up with a full football weekend slate.
The report framed Kalshi as two businesses inside one venue. The fast side carries volume. The slow side carries conviction.
Volume and holdings are concentrated in different places
Football parlays accounted for 58.0% of contract volume and held 33.0% of positions over the nine-day period. Crypto made up 14.7% of volume but just 1.6% of holdings. Election markets showed the opposite pattern: 0.2% of weekly volume, yet 24.9% of open interest.
The report’s wording was blunt. Hardly anyone trades those election markets. Everyone holds them.
Turnover data made the split even sharper. Crypto turned over 115.7 times average holdings across nine days, with average holdings of 22.33 million contracts. Football parlays turned over 34.7 times, with holdings of 294.87 million. Election markets turned over at roughly one-tenth the pace implied by their order books.
So the split was simple. The exchange’s fast half carries trading activity, while the slow half carries prediction.
Open interest hit 1.417 billion, but half the increase came from a weekly reset product
Open interest finished the period at 1.417 billion contracts, up 16.0% from 1.221 billion nine days earlier. The report argued that this was not a clean, exchange-wide record.
Football parlays are created and wiped out every week. They stood at 431 million contracts on Friday, Sept. 11, dropped to 89 million by Sunday, Sept. 13, and then climbed back to 468 million by Friday, Sept. 18.
Take parlays out and the rest of the exchange grew from 879 million contracts to 949 million, an 8.0% increase, with no weekend giveback anywhere in the series. And while the Friday close of 1.417 billion was the highest end-of-day reading in the period, the ex-parlay figure for Friday was actually below Thursday.
The September hike contract moved from 26% to 88%
Kalshi’s September rate-hike contract opened August at 56%, fell to 26% on Aug. 14, and was still at 30% on Aug. 27. It jumped to 50% the next day, just hours after Federal Reserve Chair Kevin Warsh spoke at Jackson Hole.
From there, the climb followed the incoming data. August PPI was released on Sept. 10 at +5.4% year over year, with diesel up 24.1%, and the contract closed at 63%. August CPI was released on Sept. 11 at +3.4% year over year, with gasoline up 3.9%, and the contract closed at 79%. By Sept. 15, the September hike contract had reached 88% and stayed there into the decision.
The committee then voted 12-0 to raise the target range by 25 basis points to 3.75%-4.00%. Warsh tied the move to his August remarks, saying, “As I said at the Jackson Hole policy symposium, I found it hard to describe broad financial conditions as restrictive. The committee broadly agreed with that view. So we removed one dose of accommodation.”
Two months earlier, that same committee had split 9-3, with Hammack, Kashkari and Logan dissenting in favor of a hike. Kalshi’s pricing, from 26% to 88%, tracked the resolution of that debate.
October and December pricing implies 1.22 hikes, above the committee median
The Fed released updated projections alongside the September decision. Growth, inflation and rates were all revised higher, while unemployment was revised lower by two percentage points. The median participant placed the federal funds rate at 4.1% at the end of 2026, up from 3.8% in June.
Warsh rejected the forward-guidance framing. “I do not do forward guidance,” he said. On the projections, he added: “Those are not my projections. They are the projections of my 18 colleagues.” He did not submit a projection of his own.
The report argued that the distribution below the median mattered more. The lowest of the 18 individual projections was 3.9%, the rounded new midpoint, and no participant projected a rate cut this year. Twelve of the 18 sat at 4.125%, the midpoint of the 4.00%-4.25% range, which maps to one more hike across the two remaining meetings.
Kalshi’s market priced a more hawkish path. The October contract closed at 55% and the December contract at 67%. Put together, the exchange priced 1.22 quarter-point moves across those two meetings, versus one move in the committee median.
Roughly two-thirds of the committee supported another hike by year-end, while the next meeting was priced only a little above even odds.
Little policy communication accompanied the October repricing
The report said this repricing was easier to pin down than most. The Fed was in blackout from Sept. 5 to Sept. 16. After that, the only public remarks were two speeches by Bowman on Sept. 18, and both were about bank regulation.
That means the move in October pricing from 45% to 55% happened without fresh policy communication. The other explanation offered in the report was the data released in the same period: initial jobless claims fell to 196,000 on Sept. 17, and August retail sales rose 1.2%.
The deepest order books were not in the hottest markets
On Sept. 19, the study looked across 137 Fed, CPI and Democratic nomination markets. Ten of them could absorb a $5,000 trade with no more than 2% price impact and more than $10,000 in displayed depth.
None of those ten were CPI contracts. Six were nomination markets. The October hike contract, which carried the week’s repricing, had the highest crossing cost among the surviving markets: 0.8 cents on top of a 54-cent quote.
The report said the deepest order books on the exchange belonged to a market that almost nobody traded.
After the Fed decision, Economics volume surged while holdings fell
Economics was the loudest category of the week and the biggest loser in holdings. Contract volume jumped 152.1% to 81.70 million, while holdings fell by 32.30 million.
The main reason was settlement in the September Fed contract, which traded 62.30 million contracts and saw holdings drop from 32.50 million to 1.80 million after settlement. The report said that did not signal fading interest. What people had been holding simply settled, and the capital went out with it.
The only meaningful new build in that category was KXRATECUTCOUNT, which asks how many moves occur rather than whether a move occurs. It added 1 million contracts.
What traders actually bought
KXMVECROSSCATEGORY, the football parlay ledger, traded 9.52 billion contracts, or 54.1% of all exchange volume, and finished with 362.5 million contracts in holdings. These are combined NFL and NCAA contracts settled by the leagues that run those events.
KXPRESNOMD, the Democratic nomination market, traded 6.8 million contracts, added 6.6 million in holdings, and ended at 144.6 million. The contrast in the report was stark: nearly every contract traded in the Democratic nomination market stayed in holdings, while parlays needed 26 trades to retain one contract in open interest.
Elsewhere, Super Bowl markets added 21.40 million contracts in holdings, the World Series lost 6.40 million, and the two U.S. Open singles ledgers went to zero when the tournament ended.
Perpetual futures notional reached $7.62 billion
Kalshi’s perpetual futures book traded $7.62 billion in notional value during the period, up 8.6%. The report said that was about twice the cash turnover seen in event contracts, though the two measures are not comparable: notional captures leveraged price exposure, while cash measures what traders paid for event contracts.
Inside that book, the two largest markets moved in opposite directions. Bitcoin notional rose 48.3% to $4.28 billion. Ether fell 24.3% to $2.66 billion, yet the corresponding open contracts rose 124.7%.
Lower turnover with higher holdings pointed to position building rather than active trading. The report described that as the clearest positioning shift in the perpetuals complex that week.
The product is only 16 weeks old, with gold and silver added on Sept. 10
Kalshi launched the first U.S. perpetual futures on May 29 and crossed $1 billion in its first week. On Sept. 10, it added gold and silver, the first non-crypto perpetual contracts on the platform.
Metals also triggered the only two fee-rate changes in the window. Both happened at 04:00 UTC on Sept. 10, when the market-maker fee-rate multiplier for a metal contract was set to zero, and neither change was reversed.
Silver then traded $145.96 million across 201 of the 216 hours in the period. Gold traded across 202 hours.
Newly active series were mostly sports, but retention varied sharply
A total of 513 series traded in the period after showing no activity in the prior one. The count was based on first observed trading, not the exchange’s listing date.
Most of the new activity came from sports. There were 176 sports series, led by golf. The PGA Tour ledger traded 26.76 million contracts, followed by boxing, WNBA, Europa League, AFC Champions League, and hourly temperature markets for New York and Chicago.
The report used PGA Tour and Davis Cup markets to show the broader pattern. PGA Tour retained 16.67 million of its 26.76 million traded contracts. Davis Cup markets traded 26.18 million but retained only 140,000. Similar volume. A 119 times difference in retention.
What the crowd got right, and what settles next
One forecast from the prior week had already resolved. Kalshi priced an 88% chance of a Sept. 16 Fed hike, and the Fed did raise rates on Sept. 16.
The highest-volume contracts that settled during the period were the September Fed decision, two U.S. Open singles champions, and that week’s UEFA Champions League matches.
As of the Sept. 18 close, Kalshi’s October hike contract stood at 55%, the first close above 50%, up from 31% nine days earlier. Twelve of the 18 FOMC participants still had one more hike in their projections, and two meetings remain this year. The October contract settles on Oct. 28. The December contract was priced at 67% and settles on Dec. 9.
