EKG Puts NFL Wagering at $40bn This Season With Prediction Markets Taking 21%

Eilers & Krejcik's first combined forecast credits sportsbooks with $31.7 billion and the exchanges with an $8.4 billion "handle analog", a translation of exchange volume into sportsbook terms that it has not published. The AGA sees sportsbooks flat at $29.5 billion; Week 1 data suggests the exchange share may be larger.

• Eilers & Krejcik Gaming forecasts about $40 billion of NFL wagering in 2026, comprising $31.7 billion of regulated sportsbook handle, retail and online, and $8.4 billion of prediction-market "Handle Analog", in a 16 September note reported by Gambling Insider
• That gives prediction markets about 21% of combined NFL wagering, in what EKG calls the first season "in which prediction markets compete alongside regulated sportsbooks at scale"
• The American Gaming Association projects $29.5 billion of legal sportsbook handle, flat on $29.4 billion, and blames "backdoor sports betting" on the exchanges; H2 Gambling Capital sees $31.4 billion, down 0.8%, the first decline of the legal era
• EKG's modelling suggests a "low-single-digit impact" on sportsbook handle in the most competitive states, 1% to 3%, below H2's 4% to 6%, and it is "not yet making a directional call" on substitution
• Its July data put 44% of Kalshi's retail demand in California and Texas and 69% in states without legal sports betting, a map now in flux after Missouri's letters and the Ninth Circuit's rulings

The First Forecast to Add the Two Channels Together

Eilers & Krejcik Gaming has produced its first forecast that puts regulated sportsbooks and prediction markets on one NFL scoreboard, and the only one Gambling Insider found that publishes a like-for-like split. Americans will wager roughly $40 billion on the 2026 season, the firm estimates in a 16 September note titled "NFL Wagering To Hit $40bn This Season", made up of $31.7 billion of regulated sportsbook handle and $8.4 billion of what it calls prediction-market Handle Analog, or HAN, according to Gambling Insider, which reviewed the note.

On that basis the exchanges take about 21% of combined NFL wagering and sportsbooks 79%. EKG calls 2026 the first season "in which prediction markets compete alongside regulated sportsbooks at scale" and the season "an important read on how the two channels interact", but "not a final verdict".

The sportsbook figure sits at the top of the published range. The American Gaming Association projects $29.5 billion of legal NFL handle, flat against $29.4 billion last season, and blames "backdoor sports betting on so-called 'prediction markets'" for stalling growth. H2 Gambling Capital forecasts $31.4 billion, down 0.8%, which would be the first decline in the legal era. RotoWire projects a record $32.3 billion at sportsbooks and $36.8 billion of prediction-market trading volume on NFL outcomes, a notional figure its own analysis says "does not equate to sports betting handle".

The spread reflects definitions more than disagreement. The AGA counts only commercial sportsbooks; the exchanges report notional volume that counts both sides of every trade and includes the market-maker side of long-odds parlays. EKG's contribution is a conversion from exchange volume into a sportsbook-style handle number, but it has not published the method, and Gambling Insider had not received a reply by publication to its question of whether $8.4 billion is a base case or a floor.

Week 1 is the reason to ask. Jefferies counted $3.12 billion of volume across eight exchanges on the first Sunday, of which $963 million of Kalshi's was NFL-related, and InGame put Kalshi's taker-side Sunday volume, a closer proxy for handle, at $596.7 million. EKG's own chart shows 2.94 billion football contracts traded on the exchanges it tracks from 1 to 14 September, nearly four times the 746 million a year earlier, with Kalshi up 2.5 times and Polymarket US rising from zero to about half a billion. Set against that, $8.4 billion over about 23 weeks implies roughly $365 million of NFL handle analog a week, below Needham's $2.1 billion all-sports Week 1 figure and below EKG's own $1.3 billion-a-week June run rate. On the sportsbook side, New York's record $594.6 million of mobile handle in the week to 13 September is the only hard state datapoint so far, and Gambling Insider's own extrapolation from New York's typical 15% to 16% national share puts the week's national sportsbook handle near $3.8 billion.

On substitution, EKG's modelling "suggests low-single-digit impact in the most competitive OSB states", meaning sportsbook handle 1% to 3% lower than it would otherwise be, consistent with DraftKings' claim of no discernible impact but not with BetMGM's "more meaningful negative impact", and below H2's 4% to 6%. EKG's July data put 44% of Kalshi's retail demand in California and Texas and 69% in states without legal sports betting, and it flags that if access narrows in those states, activity "could become concentrated in states where regulated sportsbooks already operate".

Handle Analog Is the Right Idea and, Unpublished, It Is an Assertion

The industry needed a number that lets a sportsbook handle and an exchange volume be compared, and EKG has supplied one. Without the method, though, 21% is a conclusion the reader cannot check. Notional volume overstates exchange activity by counting both sides and by including market-maker liquidity on parlays that no sportsbook would book at that size; taker-side volume understates it by excluding legitimate two-way retail. A conversion factor sits somewhere between, and where it sits decides whether the exchanges are a fifth of the market or, as Gambling Insider's Week 1 arithmetic and H2's estimate suggest, closer to a third. The forecast will be judged less on $40 billion than on whether the $8.4 billion turns out to have been a floor.

Sportsbooks Flat Is the Consensus, and the Argument Is About Why

Every forecaster has sportsbook handle within a few percent of last year, which for a channel that had grown every year since 2018 is the real story. The AGA says prediction markets; EKG says low single digits of that; H2 says 4% to 6%; the operators disagree with each other. What none of them has yet separated is substitution from saturation: a market that has legalised most of the states it will legalise, and whose calendar this September contains fewer NFL games than last, would have flattened anyway. The exchanges are the convenient explanation, and they may be the correct one, but the evidence for the size of the effect is the same Week 1 that the GeoComply data showed producing record sportsbook activity.

The Map Is the Variable EKG Is Right to Flag

Sixty-nine per cent of Kalshi's retail demand in states without legal sports betting is the figure that matters for substitution. If that demand is incremental, the exchanges are growing the pool; if enforcement pushes them out of those states, the remaining demand sits in states where a sportsbook was available and the exchange was chosen anyway, and the substitution rate rises. Missouri's letters, Connecticut's, Washington's injunction and the Ninth Circuit's two rulings are all pushing in that direction inside the same season EKG is measuring.

EKG has put a share on the exchanges and a caveat on the share. The season will decide which of the two the industry remembers.