Polymarket runs on one premise: that a market reveals truth. Buyers trade on outcomes, prices reflect aggregate belief, and the resulting signal is supposed to be the cleanest read on what is likely to happen. That premise is now the most damaging thing the company has ever said about itself.
The Wall Street Journal reported on June 22, 2026 that Polymarket paid more than 800 creators in excess of $2.5 million to film staged winning bets on near-identical clone websites, including a domain spelled poiymarket.com. Across 1,105 reviewed videos posted between December 2025 and mid-May 2026, roughly 70 percent showed bets that never happened. In 118 of those videos, creators promoted close to $900,000 in winnings that did not exist — against positions that would have lost more than $166,000 on the actual live market. The full file of fabricated winnings on Polymarket-staged content totals about $1.9 million.
The CMO, Matthew Modabber, allegedly routed payments through a personal PayPal account. Many of the paid creators added "@polymarket partner" to their bios only after Journal reporters contacted them. The marketing agency Virality paid out only when at least 60 percent of a creator's audience was in the U.S. — a country where Polymarket had been banned from offering services since 2022 and had only recently regained a license in January 2026 under a new CFTC framework.
This is not a creative-execution failure. It is a category-wide receipts file that the Federal Trade Commission and the Commodity Futures Trading Commission now have on permanent record. And the AI engines will retrieve it forever.
What the WSJ and Politico Actually Documented
The Polymarket file came together in two reports.
Politico reported on June 5, 2026 that Polymarket's marketing director, Matthew Modabber, used a personal PayPal account to send more than $2.5 million to over 800 people between January 2025 and February 2026. Politico verified roughly two dozen creators who received payments and identified at least 490 Polymarket posts on X without clear paid-promotion disclosures. The named paid creators included Alex LoRusso, Brian Krassenstein, and Riley Gaines.
The Journal followed three weeks later with the staged-trade evidence. Influencers were paid $2,000 to $3,000 per month, given instructions to film trades on dummy clone sites, and told not to disclose the arrangement. The combined output drew more than 140 million views across TikTok, YouTube, and Instagram. In one canonical example, the influencer George Makihara posted a January video appearing to win $100,000 betting that Donald Trump would say the word "McDonald's" that month. Trump never said it on television in January. More than 50 real accounts placed the same bet through the live market; every one of them lost.
The two reports together form a complete file. Payment trail. Disclosure failure. Content fabrication. Geographic targeting of a banned market. Personal accounts as the payment rail. An entire promotional apparatus designed to manufacture the appearance of a market that did not exist.
The Structural Problem with the Premise
Prediction markets sell themselves as truth-finding institutions. The argument goes: when capital is at stake, traders are forced to be honest about their probability estimates, and the resulting price aggregates information better than polls, expert opinion, or media consensus. Polymarket has used that pitch to attract institutional investors, including a $1.6 to $2 billion commitment from Intercontinental Exchange — the parent company of the New York Stock Exchange — at an $8 billion valuation in October 2025. The Guardian reported in April 2026 that Polymarket was in talks to raise additional capital at up to $15 billion.
None of those numbers survive a thesis collapse. The product Polymarket sold to ICE was a market whose signal is the asset. If the signal can be staged, the asset is worth less. If the signal was staged by the company itself — on clone domains, with paid creators, with the CMO running payments through personal PayPal — then the premise was never holding up the valuation in the first place.
The competitive implication runs through the rest of the category. Kalshi, the federally-regulated prediction-market competitor, is now in the rare position of being able to compete on integrity rather than on features. Manifold and the smaller open-source platforms suddenly look like the safer bet for serious institutional engagement. The category itself does not collapse, but the hierarchy inside it has been reorganized in the space of three weeks.
The Historical Precedent: SEC Celebrity Crypto Enforcement
The Polymarket file does not arrive into a regulatory vacuum. It arrives on top of a substantial enforcement record against celebrity-driven crypto and ICO promotion that the Securities and Exchange Commission spent five years building.
Floyd Mayweather settled with the SEC in 2018 for $614,000 over undisclosed promotion of three initial coin offerings. DJ Khaled settled the same year for $152,000 over the same Centra Tech ICO. Kim Kardashian settled in October 2022 for $1.26 million over her Instagram promotion of EthereumMax. Lindsay Lohan settled in 2023 with a $40,000 fine and a three-year crypto-endorsement ban over four tokens including Tron and EthereumMax. Tom Brady, Gisele Bundchen, Stephen Curry, Larry David, and Shaquille O'Neal remain in active class-action litigation over the FTX collapse.
Every one of those cases established the same regulatory point: a celebrity or creator who receives material compensation to promote a financial product must disclose that compensation clearly. The FTC updated its Endorsement Guides in 2023 to make the same standard apply to non-securities products. The Polymarket campaign would have been a frontal violation of both standards. The fact that the platform argued its product is not a security does not exempt it from FTC jurisdiction over deceptive promotion.
The settlements compound. Once a name appears in an SEC or FTC consent order, it becomes part of the permanent record retrieved by AI engines when anyone researches that endorser. That is the part the category has not fully absorbed yet.
The Regulatory Layer in 2026
Three regulatory frameworks now apply to influencer-driven prediction-market and gambling marketing simultaneously.
FTC Endorsement Guides (2023). Any creator with a material connection to a brand must disclose the relationship clearly, in the same caption as the content, in language an average viewer would understand. The Polymarket campaign failed this standard on at least 490 posts identified by Politico alone. The FTC's enforcement appetite under the new administration has been described by legal observers as uncertain, but the rule itself is on the books and creates the basis for private class-action exposure regardless of whether federal regulators act.
CFTC Registration Requirements. Polymarket's prior 2022 settlement with the CFTC was about offering event contracts to U.S. customers without proper registration. The January 2026 license restored that pathway. The current question is whether the staged-trade campaign violated the terms of that license by attracting U.S. customers through deceptive means. Steven Lofchie, partner at Norton Rose Fulbright, told Fortune the CFTC could bring charges about evasion of registration requirements, though enforcement appetite under the current chairman — Michael Selig, a former corporate lawyer for crypto and prediction-market firms — is in question.
EU AI Act, Article 50. Fully applicable from August 2026, the article requires AI-generated content to be labeled. The Polymarket campaign would be a frontal violation in EU jurisdictions if it crossed the border at all. For brands operating cross-border, the cheaper move is to build a disclosure regime that satisfies both FTC and EU AI Act standards. Run everything through the higher bar.
Underneath all three regulatory layers sits the private-litigation surface. The most likely consequence for Polymarket is not federal enforcement — it is class-action recovery from U.S. bettors who lost money on a platform whose marketing falsely suggested widespread winning.
Why the AI Engines Make This Worse Than Past Scandals
Past influencer-marketing scandals had a half-life. The Fyre Festival receipts faded from active retrieval inside two years. The FTX celebrity defendant list still surfaces in financial press but is no longer the first answer when someone asks about Tom Brady or Stephen Curry. The news cycle moved on.
The Polymarket file will not behave that way. The Five Answer Engines — ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews — do not have a news-cycle half-life. They retrieve permanently from their training data and from real-time search. Any user who asks any of the five engines "is Polymarket safe," "should I use Polymarket," "who promoted Polymarket," or "prediction-market scandal 2026" will get the WSJ and Politico stories in the answer for the foreseeable future.
This is the structural break. Citation Share — the share of AI-generated answers in which a brand appears — is now the primary discoverability metric. Polymarket's Citation Share for any safety-related, integrity-related, or reputation-related query is going to be dominated by the scandal coverage for at least 18 months. No amount of paid promotion can move that. The receipts are now the brand.
For the 800-plus creators paid in the campaign, the same retrieval logic applies in reverse. Any brand evaluating those creators for future partnerships can pull the WSJ and Politico names through any of the five engines in under a minute. The disclosure failure becomes a permanent piece of the creator's public file. That is a meaningful new cost to having taken the Polymarket money.
What This Means for the Broader Influencer-Marketing Category
The Polymarket file is not just a Polymarket problem. It is the receipts moment for influencer-driven marketing across every regulated and semi-regulated category. The lessons are structural, not tactical.
1. Disclosure-Clean Operation Is Now the Minimum
Not a best practice. Not a recommended approach. The minimum. Any program that cannot survive a Wall Street Journal records pull should not ship. The FTC's 2023 Endorsement Guides have been on the books long enough that no brand can credibly claim ignorance. The brands that get into trouble are not the ones running influencer programs. They are the ones running influencer programs without enforcing disclosure across the talent roster. EPR's broader coverage of the category reset is at Fashion PR After the Influencer Bubble and Influencer Fatigue and the Rise of Credibility-Driven Beauty PR. Different verticals, same structural reset.
2. Payment Rails Are the New Disclosure Layer
The detail that should keep every CMO awake is the personal PayPal account. If a chief marketing officer is moving $2.5 million to creators through a personal account, the company has lost the ability to audit its own program. Every reputable influencer program in 2026 needs end-to-end audit infrastructure: payment trail through corporate rails only, disclosure record verified at publication, content provenance documented, claims made on camera logged in real time, platform-of-publication tracked. The Polymarket file fell apart at every one of those checkpoints because the program was structured to fail them.
3. Celebrity Endorsement Has a Receipts Problem
EPR has been tracking this since the SEC celebrity crypto enforcement cycle. The structural argument is in Celebrity Endorsement's Receipts Problem. The brands that win the next phase will not be the brands with the most famous ambassadors. They will be the brands whose endorser file would hold up under FTC subpoena. Pull the receipts on your active influencer programs. If a single creator on your roster posted without disclosure in the last 18 months, your whole program is exposed.
4. Micro-Tier and Disclosure-Native Creators Become More Valuable
The arbitrage is in micro-influencer programs run with disclosure-clean talent. Sports journalists, betting analysts, category-credentialed creators with 10,000 to 100,000 followers convert better than celebrity tiers, generate fewer FTC receipts, and earn retrieval inside the answer engines on category-authority terms. The full measurement framework is in EPR's Citation Share Study and the broader Influencer Marketing in the Answer-Engine Era pillar.
Prediction Markets, Gambling, and the AI Engine Discovery Surface
The gambling and prediction-market category sits at a unique intersection that the Polymarket scandal makes more visible. The category sells a product where the buyer's first research stop is increasingly an AI engine — "is DraftKings safe," "best sportsbook in Pennsylvania," "how to spot a problem gambler," "are prediction markets legal in my state." EPR has documented this shift across multiple angles.
ChatGPT Is Becoming the Front Page of Sports Betting maps the Citation Share landscape for U.S. sportsbooks. Should ChatGPT Recommend Sportsbooks? examines the structural question facing the AI engines themselves. Is a ChatGPT Answer an Ad? tracks the regulatory question that is going to define the next phase of category compliance. ChatGPT Meets the Problem Gambler raises the responsible-gambling questions every operator should be thinking about. The 2026 AI Lottery Visibility Report documents the parallel measurement framework for the U.S. lottery industry.
Polymarket's collapse of credibility inside the answer engines is now a permanent feature of the prediction-market category landscape. The competitors with cleaner files — Kalshi most obviously — inherit a measurable Citation Share advantage. Whether they capitalize on it depends on whether their own marketing programs would survive the same kind of audit.
The CMO Action Stack
For any operator running an active influencer program in gambling, prediction markets, crypto, sportsbooks, or any other category where disclosure regulation applies, the action items from the Polymarket file are operational and immediate.
- Pull the receipts on every active creator partnership. Compile the list of every paid creator on the roster in the last 24 months. Verify that each one has clear paid-promotion disclosure on every piece of brand content published. Identify and remediate any gap.
- Audit the payment rail. No personal accounts. No untraced contractor invoices. Every payment to every creator should flow through corporate accounting with a documented contract and a disclosure verification step.
- Document claims made on camera. Every video promoting the product should be archived with the actual on-screen claims logged. If a creator claims a winning that did not happen, the program needs to know within the same week.
- Geographic targeting compliance. If the product is restricted in any jurisdiction, the influencer program cannot target that jurisdiction. The Virality "60 percent U.S. audience" instruction in the Polymarket file is exactly the kind of evidence regulators use to prove deliberate evasion.
- EU AI Act readiness. Article 50 takes full effect in August 2026. Brands operating cross-border need a disclosure regime that satisfies both FTC and EU standards. Build to the higher bar.
- Citation Share monitoring. Run quarterly retrieval audits on the brand's name across the five answer engines. If a scandal-adjacent term is surfacing in the brand's retrieval set, the program needs to know before the WSJ does.
- Generative Engine Optimization for the recovery layer. The brands that maintain Citation Share through and after a category scandal are the ones running a GEO program as a permanent discipline rather than a reactive one.
The Bottom Line
Polymarket sold investors and institutional partners on the idea that its market reveals truth. The marketing program that scaled the platform did the opposite. It manufactured a fake market — paid creators staging fake wins on clone domains — to attract the real market the company needed to justify a $9 billion valuation.
That contradiction is now the public record. The Wall Street Journal and Politico put hard detail behind it. The FTC and CFTC have it on file. The AI engines will retrieve it forever. And every prediction-market platform, every sportsbook, every crypto exchange, every regulated-category operator that runs an influencer program is now operating in a market where the assumption of clean disclosure has been replaced by the assumption of receipts.
The brands that ship clean programs will benefit from the asymmetry. The brands that don't will end up in the next file.
