$34 billion. That is the number cited as the volume US-based users moved through offshore, unlicensed prediction-market venues during a recent twelve-month window, according to a study now circulating across English-language financial press and most of the trade-publication aggregators that follow this beat. The conventional reading is that the figure is the regulated sportsbook industry's most damaging public-relations event of 2026, and the strongest single argument for federalising event-market regulation before the offshore venues become too entrenched to disrupt.
The framing has weight. US online sports betting was a $13.7bn market in 2024 on the public record, per the consolidated disclosures filed by Flutter Entertainment at its investor results centre. If the $34bn offshore figure is even approximately accurate, the unlicensed market is running roughly two and a half times the size of the regulated one. That is not a leak. That is a parallel industry, sitting alongside the licensed product with a meaningfully larger top-line and a meaningfully narrower compliance perimeter.
Add the political-event dimension — contracts written on election outcomes, Federal Reserve decisions, Supreme Court rulings, geopolitical resolutions — which the regulated US sportsbooks cannot legally offer, and the case crystallises into the conventional reading the trade press has settled on. The offshore venues are not simply absorbing sports bets that should have routed to FanDuel and DraftKings. They are absorbing an entire product category the current US licensing framework refuses to license at all.
Why This Is Actually True
We will concede the structural case fully because the structural case is sound. FanDuel publishes 22 legal US states for its sportsbook product on its own domain disclosures. DraftKings publishes 27. Both are properly licensed in Ontario through AGCO and have full New Jersey NJDGE permits. Neither operator services Texas, California, Florida, or Georgia for online sports wagering. Roughly half the US population sits outside the legal sportsbook perimeter as a matter of state law, and inside that excluded population the demand to express a view on a sporting outcome does not vanish. It re-routes.
Once it re-routes, it does not necessarily come back. The same retail behaviour that ports a sports bet to an offshore venue ports a Fed-decision contract or an election-outcome contract to the same venue. The product surface widens. The user gets used to the deposit rail. The CRM relationship deepens. Five years of compounding behaviour produces exactly the offshore-platform stickiness the regulated industry is now publicly worried about.
The numbers the regulated US operators do disclose make the gap visible. Flutter's US segment booked $6,180m of revenue in 2024. DraftKings booked $4,770m. The combined regulated-US licensed revenue base, even before subtracting BetMGM's $26-state footprint and the various state-monopoly operators, sits at roughly $11bn. A $34bn unlicensed shadow market, if real, is sized to dwarf that licensed revenue base by three to one. The structural conventional wisdom — that US event-market regulation has produced an arbitrage too large for licensed operators to compete with on volume — is not a marketing fiction. It is the math.
But the $34bn figure is a study citation. The study itself is what this desk cannot pull into the dataset, and that gap is where the editorial lives.
Where It Breaks Down
Rule 1 of this publication is that every operator claim, every revenue figure, every market-size estimate is grounded against the primary document or it is flagged. We cannot pull the originating study for the $34bn figure into our dataset. The number is widely cited; we are not in a position to verify it against the methodology document, the sampling approach, the wallet-tracing logic, the offshore-venue identification list, or the volume-versus-handle definitional question that decides whether $34bn measures gross transaction throughput or net economic activity. The study may be sound. It may be approximately right. It may be off by a factor of two in either direction. We do not know, because the dataset behind it is not in front of us.
What we can verify, against the Flutter investor disclosures, is that the same Flutter filing which sources the $13.7bn US online sports betting market estimate also discloses FanDuel as holding 43% US sportsbook market share and contributing 44% of group revenue. Those are precise figures with audit-committee liability attached. The $34bn comparator does not carry equivalent liability. Comparing one against the other as if both numbers occupy the same epistemological tier produces a misleading ratio.
Two definitional questions also do most of the work in the headline framing. First, does $34bn cover only sports-event contracts traded on offshore prediction-market platforms, or does it bundle political contracts, weather contracts, crypto-price contracts, and entertainment-industry contracts into one undifferentiated total? Second, is the figure gross transaction value or net of contract closeout and counterparty round-tripping? On most prediction-market venues, the same dollar of user capital can transact across both sides of a binary contract several times before settlement. Gross volume is not handle. Handle is not gross gaming revenue. The conflation matters when the resulting figure is used as a comparator to a regulated GGR number.
A reader who relies on the conventional headline framing without unpacking either question is making policy or investment inference from a study they have not read against filings they have not opened.
The Rule I Use Instead
Read the filing, not the headline. When a regulated operator discloses a market-size number, the disclosure is bounded by the audit liability of misrepresenting it to securities regulators. When a study cites a market-size number, the study has no equivalent audit liability and the methodology often does not survive the kind of footnote-scrubbing a hedge-fund desk applies to a 10-K. Both numbers can be cited. Only one is investable, and only one is the basis for an enforcement action against the issuer if it turns out to be wrong.
Entain's annual report for 2024 is the parallel example that demonstrates the discipline. The filing discloses regulated-markets revenue at 88% of group consolidated revenue, with the remaining 12% explicitly tagged as gray-market exposure and broken out in the segment commentary the analysts on the call were expected to query. Group revenue is £4,833m. Active customers, 28 million. Those numbers are reconciled, audited, and tied to the prior-year DPA disclosure on the Turkey-facing legacy business that settled at £585m with the UK CPS. An analyst reading that filing closely can build a precise map of where the regulated economic activity sits and where the residual unregulated exposure sits, and can do so against figures the issuer is legally obligated to defend.
The same discipline applied to the $34bn prediction-market claim asks: which jurisdictions hold the offshore venues' enforcement records, what is the methodology for wallet attribution to US-based users, and which licensed alternative actually exists in the same product category. Without those answers, the figure is a directional signal rather than a comparator. Directional signals belong in opinion columns. They do not belong in policy testimony, and they do not belong in regulated-industry investor presentations as if they were settled fact.
The reading rule generalises beyond this one number. Numbers cited from filings carry the weight of the filings. Numbers cited from studies carry the weight of the studies. The two weights are not the same.
When the Old Rule Still Wins
Here is the concession. The offshore prediction-market phenomenon is real even if the $34bn headline figure is unverified by this desk. Roughly half the US states do not license online sports betting at all on the public register of operator coverage for any equivalent licensed product, and zero US states license political-event contracts in any form a retail user can access. That gap will continue producing offshore demand regardless of how the headline figure is sourced. The licensed responsible-gambling infrastructure most readers of this publication take for granted — the GAMSTOP self-exclusion register that binds every UKGC-licensed online operator automatically, the OASIS cross-operator deposit cap enforcement in Germany, the segregated player-fund disclosures in the regulated operator 10-Ks — has no counterpart on the offshore prediction-market venues. A user who routes their event-market activity offshore has stepped outside the entire consumer-protection infrastructure that the regulated jurisdictions spent fifteen years building.
The structural case for federalising US event-market regulation does not need the $34bn figure to be exactly right. The case stands on the regulatory gap itself, which is on the public record, in the state-by-state licensing tables every operator publishes for compliance.
FAQ
Why does this desk flag the $34bn figure rather than just citing it?
Because the originating study is not in our verification dataset. Our editorial discipline is to pull every cited number back to the primary document — the filing, the regulator notice, the audited report — and flag the gap when we cannot. The $34bn estimate may well be correct. We are not in a position to confirm the methodology, the sampling, the wallet-attribution logic, or the definitional question of whether the figure is gross transaction volume or economic activity. Citing it as settled fact without that confirmation would breach the publication's grounding rule.
How big is the regulated US online sports betting market by comparison?
US online sports betting was a $13.7bn market in 2024 according to the consolidated disclosure in Flutter Entertainment's annual investor materials. FanDuel held 43% of that market by Flutter's own segment reporting. DraftKings disclosed $4,770m of revenue in fiscal 2024 on its own filings. The licensed market is sized, audited, and on the public record. The figures carry securities-law liability if misrepresented.
Does the gap between $13.7bn licensed and $34bn offshore mean the offshore market is bigger?
Only if both numbers measure the same thing, which is the open question. The $13.7bn figure is regulated US online sports betting market size. The $34bn figure, as cited, appears to cover a wider product surface that includes political contracts, weather contracts, and other event categories that are not licensed in any US state for sportsbook operators. Apples-to-apples comparison would require restricting both numbers to the sports-betting category and reconciling gross transaction volume against net economic activity.
Why can't licensed US sportsbooks just offer political event contracts?
State sportsbook licensing frameworks are scoped to sporting events as defined by the relevant state regulator. Event contracts on elections, Federal Reserve decisions, or Supreme Court rulings sit in a regulatory category overseen federally rather than at the state level, and the licensed sportsbook operators are not currently authorised to offer them under the state permits they hold. The product category is structurally absent from the licensed surface, which is the precise gap the offshore venues serve.
Where does this leave consumer-protection coverage for US users on offshore venues?
Outside it. The responsible-gambling infrastructure that binds UKGC, MGA, AGCO Ontario, and NJDGE licensees — single-registration self-exclusion via GAMSTOP, cross-operator deposit caps under the German GGL framework, segregated player-fund disclosure requirements in the regulated 10-Ks — does not extend to offshore prediction-market venues. A US user who routes activity offshore has stepped outside fifteen years of accumulated consumer-protection regulation, and there is no equivalent backstop on the destination side.
What would make the $34bn figure investible rather than directional?
A published methodology document with sampling logic, a defined offshore-venue list, a wallet-attribution approach that survives independent replication, and an explicit definitional position on gross volume versus economic activity. Audited issuer disclosures meet those criteria. Standalone studies sometimes do and sometimes do not. Until the underlying document is on the public record in a form that allows an analyst to replicate the calculation, the figure functions as a directional signal rather than a comparator to a regulated GGR number.