The Price of Truth — Working Hypothesis

On January 7, 2026, Karoline Leavitt’s White House press briefing ran roughly sixty-four and a half minutes.

For most people watching, it was unremarkable. For a subset of Polymarket and Kalshi traders, it was a disaster. They had bet — at 98% implied odds — that the briefing would hit sixty-five minutes. When it didn’t, they erupted on social media. The White House press secretary hadn’t known she was racing a clock. The traders had turned a government information event into a speculative contract, and they were furious the outcome hadn’t cooperated.

A month earlier, Kalshi had announced a partnership with CNN. The briefing incident is what that partnership looks like in practice.

What the consensus gets right

Prediction markets have a genuine intellectual history. The core argument — that people with financial skin in the game will aggregate information more accurately than polls, pundits, or surveys — goes back at least to Friedrich Hayek’s price theory, and was endorsed in 2008 by a letter from twenty-two prominent social scientists including four Nobel Prize winners in economics.

The 2024 election gave this argument its most publicized proof point. While polling averages were split, Kalshi and Polymarket both had Trump’s probability significantly above 50% in the final weeks of the race — reaching 66% on Polymarket by late October. When the results came in, the prediction market crowd declared victory.

There’s something real here. Markets do react faster than polls. When Biden’s debate performance collapsed in June 2024, prediction market prices moved within hours while traditional polling took days to catch up. The skin-in-the-game logic is sound in theory: if you’re wrong about something with money on the line, you lose. That creates incentives for accuracy that asking someone their opinion over the phone does not.

The consensus is correct that prediction markets provide a faster, more continuously updated signal than traditional methods. It’s asking the wrong question about what that signal actually is.

What the consensus misses

Tarek Mansour, the CEO of Kalshi, gave a talk at the Citadel Securities Future of Global Markets conference in October 2025. When asked what it would take for his platform to become a trillion-dollar asset class, he said: “The long-term vision is to financialize everything and create a tradeable asset out of any difference in opinion.”

Kalshi, he noted, means “everything” in Arabic.

This is not a truth machine. This is a revenue model. And once you see it as a revenue model, the recent wave of media partnerships stops looking like journalism infrastructure and starts looking like what it is: a customer acquisition strategy disguised as editorial credibility.

Here’s what happened between election day 2024 and December 2025. At the peak of the 2024 election cycle, Kalshi had around 400,000 daily active users. Polymarket had around 300,000. By June 2025, both platforms had lost more than 90% of that activity — down to 27,000-32,000 and 5,000-10,000 daily users respectively. The election was the product. After the election, the product was gone.

Then, between November 2025 and February 2026, in the space of roughly four months, the following deals were announced: CNN with Kalshi, CNBC with Kalshi, Dow Jones with Polymarket, Yahoo Finance with Polymarket, Substack with Polymarket. Fox News is reportedly in advanced talks. The financial terms of every single deal are undisclosed.

According to a New Yorker report, Kalshi is paying CNN. CNN’s editorial employees are prohibited from trading on Kalshi. CNN’s chief data analyst is now the primary vehicle for broadcasting Kalshi’s price signals to a national television audience. A live ticker runs odds on cabinet departures, Fed decisions, and election outcomes during news segments.

This is not a data partnership. The data is the product Kalshi is selling. CNN is the storefront.

The accuracy premise

The media deals only work if the “truth machine” narrative holds. So it’s worth looking at what the actual evidence says about Kalshi’s accuracy.

A September 2025 working paper from researchers at University College Dublin — analyzing over 313,000 Kalshi contracts — found that prices are informative and improve as events approach. That’s the good news. The paper also found a systematic favorite-longshot bias: contracts priced at 5 cents win only about 2% of the time, losing 60% of invested capital. The average pre-fee return across all contracts is minus 20%. Even before Kalshi takes its cut, the aggregate player loses money at a rate that implies systematic mispricing on the low-probability end of the market.

A separate Vanderbilt study analyzing 2,500 markets with $2.5 billion in volume found Kalshi hit 78% accuracy on political markets — better than Polymarket’s 67%, but measured against a benchmark that includes base rate prediction. Kalshi’s spokesperson responded that the methodology “completely misunderstands prediction markets.”

Neither of these findings destroys the accuracy claim. Markets can be biased in one direction and still useful as signals. But they do complicate it. When CNN runs a Kalshi ticker showing “19% chance Pam Bondi is the first cabinet member to leave” during a news segment, the viewer is being given a probability estimate. The research says that estimate carries a systematic bias that its producers have not disclosed and may not have examined.

The consensus treats the 2024 election accuracy as the proof. What it skips is that the 2022 midterms were a different story. An Asterisk Magazine analysis found that prediction markets underperformed expert forecasters like FiveThirtyEight on key Senate races — giving Democrats roughly 32% odds for Senate control versus FiveThirtyEight’s 41-50%. Yale’s financial markets researchers concluded the track record didn’t support the hype. The 2024 success is also complicated by documented large-position traders on Polymarket — at least one French national operating through four coordinated accounts who made roughly $85 million betting on Trump — whose capital weight influenced the prices that were then reported as consensus.

The working hypothesis: Kalshi’s media partnerships — CNN, CNBC, Dow Jones, all executed in a four-month window after user counts fell more than 90% from their election-cycle peak — are not journalism deals. They are customer acquisition arrangements, undisclosed in their financial terms, that convert a contested accuracy claim into institutional legitimacy, administered under a regulatory framework being reshaped by a former Kalshi board member.

The regulatory architecture

In November 2020, Kalshi was approved by the CFTC (Commodity Futures Trading Commission) as a Designated Contract Market — the same category as the Chicago Mercantile Exchange. That classification is load-bearing. It means:

Kalshi is regulated as a commodity derivatives exchange, not a gambling operation
Federal law preempts state gambling regulations under the Supremacy Clause
The CFTC is the exclusive regulator — not state gaming commissions in Nevada, New Jersey, or Arizona

This is why Kalshi could tell Nevada it had no authority over its sports contracts and win in federal court. The legal argument is that a Kalshi NFL contract is structurally equivalent to a CME oil futures contract.

Under the Biden administration, the CFTC said election contracts on Kalshi were unlawful and fought that argument in court. Kalshi won at the district level in September 2024. The CFTC appealed. In May 2025, under the Trump administration, the CFTC dropped the appeal entirely.

Brian Quintenz, who served on Kalshi’s board, was nominated by President Trump to chair the CFTC in February 2025. Quintenz’s nomination was withdrawn on September 30, 2025, after stalling in the Senate Agriculture Committee amid conflict-of-interest scrutiny over his Kalshi ties. Trump then nominated Michael Selig, who was confirmed by the Senate on December 18, 2025 and sworn in as the 16th CFTC Chairman on December 22. Selig has described his posture toward prediction markets as “pro-innovation” and has withdrawn the proposed rule that would have prohibited political and sports contracts. He is now writing new guidance.

The regulatory clock that stood between Kalshi and a national media distribution deal ran out in May 2025. The media deals followed in November.

The business model you’re not supposed to notice

Kalshi’s public disclosures show that sports betting constitutes 89% of fee revenue — $235 million of $263.5 million in 2025 — with the platform processing over $1 billion in NFL volume in its first month of sports contracts alone. It is, in operational terms, DraftKings with a better lawyer.

The political and geopolitical markets — the ones CNN reports on — are economically marginal. They exist not because they generate revenue but because they generate legitimacy. Every time a CNN anchor reads a Kalshi probability on a Senate race or a cabinet departure, it tells a potential customer that Kalshi is financial infrastructure, not a betting app. That reframing is worth far more to the business than any transaction fee on a Fed decision contract.

Polymarket’s CEO Shayne Coplan told 60 Minutes that his platform was “the most accurate thing we have as mankind right now.” There is no serious evidentiary basis for this claim. What there is: a $9 billion valuation, a NYSE parent company investor via ICE’s (Intercontinental Exchange) initial $1 billion stake — part of a commitment of up to $2 billion — and a wave of media deals that circulate the accuracy claim to audiences who won’t look up the Vanderbilt study.

The Polymarket CEO also told Axios in November 2024 that it was “super cool” that his platform creates a financial incentive for people to divulge information to the market — including, he said, potential insiders.

He meant this as a selling point.

What this changes

None of this means prediction markets are useless. They’re not. The Federal Reserve published a paper in early 2026 finding that Kalshi’s macroeconomic contracts — on CPI releases, Fed rate decisions, GDP — achieve accuracy that rivals professional forecasting surveys. The skin-in-the-game argument works in domains where the relevant information is dispersed across many participants who have actual expertise and actual money on the line.

The problem is not prediction markets as a forecasting tool. The problem is prediction markets as a journalism product — specifically, as a journalism product whose commercial terms are undisclosed, whose accuracy properties are overstated in the editorial framing that accompanies the deal, and whose regulatory protection was obtained through a personnel decision with a direct conflict of interest at its center.

When Karoline Leavitt’s briefing ended 30 seconds short and traders erupted, they weren’t reacting to a forecasting failure. They were reacting to a world not cooperating with a financial position. That’s what the CEO means by “financialize everything.” Every government event, every press briefing, every cabinet departure is a binary contract waiting to happen. The newsroom is what converts the contract into a truth claim.

The media organizations in these deals are not just providing distribution. They are providing the one thing prediction markets need most: the institutional credibility that allows a gambling platform’s prices to be reported as probability.

Part 1 of 2. Next: The Classified Alpha — the same regulatory framework that made the CNN deal possible also created a financial incentive structure that rewards anyone with nonpublic information about future events. Including people with security clearances.

What would change my mind

If CNN discloses the financial terms of its Kalshi deal and they are purely informational — no payment to CNN, no revenue share, no undisclosed equity — the “customer acquisition dressed as journalism” argument weakens significantly. The conflict of interest changes shape if the money flows differently than reported.
If Kalshi’s political market accuracy — specifically the contracts CNN reports on — is demonstrated to be materially better calibrated than the overall sample from the UCD study — meaning the favorite-longshot bias doesn’t apply to the high-profile, high-liquidity political contracts — then the accuracy premise strengthens and the media framing is more defensible.
If the CFTC’s new rulemaking under Selig produces mandatory disclosure requirements for media partnerships — requiring CNN, Dow Jones, and others to disclose financial terms and accuracy limitations alongside any reported prediction market data — then the information gap narrows and the legitimacy arbitrage becomes harder to run.

If you found this useful, the best thing you can do is forward it to one person who would push back on it. I’d rather be wrong in public than right in private.