The Reveal Died at 4:32 PM: How Prediction Markets Ate the Earnings Calendar, Looking Back From 2036

The quarterly earnings call was never an information event. It was a scheduling convention — a fixed time at which a company agreed to stop withholding what it already knew. The number existed before the call. The call just decided when you were allowed to react to it. Strip away the ritual and the earnings calendar was a queue, and queues are only useful when information is scarce enough to ration.

From the vantage of 2036, the thing that killed the reveal looks obvious. Prediction markets did not get better at guessing earnings. They got rid of the queue.

Here is the conventional wisdom from a decade ago, and why it was wrong. The story everyone told in the mid-2020s was that prediction markets were forecasting tools — Las Vegas with a CFTC license, useful for elections and Fed odds and not much else. The skeptics said the volume was too thin to matter, the markets too easily gamed, the regulators too hostile. They were wrong on all three counts, and the order in which they were wrong is the whole story.

Force one: liquidity arrived, and with it, legitimacy. The thin-volume objection died first. In 2024, Kalshi and Polymarket together cleared a few billion dollars a month, most of it political — Polymarket alone took in over $3 billion on the U.S. presidential race.¹ By April 2026, combined monthly volume on the two platforms had climbed to roughly $24 billion, up from under $5 billion the previous September, with sports and crypto carrying the load that politics once did.² Kalshi raised at an $11 billion valuation in late 2025 and was reportedly in talks at far higher marks by mid-2026, having struck data deals to pipe its real-time probabilities into CNBC and CNN.³ When a forecasting instrument becomes a chyron on Squawk Box, it has stopped being a novelty and started being infrastructure. This is the law of enhancement: prediction markets amplified the one thing markets are theoretically best at — aggregating dispersed information into a single price — and did it continuously, in public, at a refresh rate the earnings calendar could never match.

Force two: the legal question resolved in the markets' favor. The "regulators will kill it" objection died next. When the CFTC tried to block Kalshi's election contracts, the D.C. Circuit declined to stay a lower-court ruling against the agency, and the CFTC ultimately dropped the fight in 2025.⁴ Event contracts were not gaming; they were a regulated financial product. That settled the franchise. Academic finance had argued for years that prediction markets "quickly incorporate new information, are largely efficient, and impervious to manipulation," and that they generally beat professional forecasters and polls.⁵ The attempts to rig sentiment that everyone feared turned out to be self-limiting, exactly as the efficiency literature predicted: a market deep enough to move is a market deep enough to punish the manipulator. This is obsolescence. What got pushed aside was not the analyst — it was the calendar. Once a liquid, legal, continuously-priced market existed for "will this company miss guidance," the discrete announcement stopped being the moment that mattered. The number leaked into the price before the CFO reached the podium.

What the prediction markets retrieved is the older thing hiding underneath all of this: the betting market as a forecasting institution. Before scientific polling, newspapers printed gambling odds on elections daily; market prices were the original forecast.⁶ The twentieth century buried that under polls and quarterly disclosure and the priesthood of sell-side research. The 2030s dug it back up and wired it to a fiber-optic backbone. Price discovery returned to its oldest form — a crowd betting real money on an outcome — now running every second instead of every quarter.

Now the reversal, which is where the whole thing turns on itself.

The promise was efficiency: perfect information, perfectly distributed, instantly priced. The reversal is that a system optimized for information efficiency optimizes against its own stability. To price one company's quarter continuously, the markets had to link it to everything that might move it — supplier health, a strait's geopolitics, a chip shortage, a weather contract in a farm belt. Each link was rational. The sum was a single, tightly coupled machine. And tightly coupled systems are the ones that fail in ways no operator can stop: this is the core of Charles Perrow's Normal Accidents, the finding that complexity plus tight coupling makes cascading failure not a bug but an inevitable property of the system.⁷ The markets did not abolish George Soros's reflexivity — the feedback loop in which beliefs move prices and prices move beliefs — they industrialized it.⁸ Sentiment now propagates at the speed of the data feed. A probability shift three derivatives removed from a company's fundamentals can reprice the company before any human reads the underlying news. Diversification assumed assets move independently; coupling made them move together. The tool built to make markets legible reversed into a machine that makes them harder to live inside.

That is the durable lesson, and it is not really about earnings calls. Information efficiency and system stability are different goods, and past a certain point you trade one for the other. Every market we have ever built to remove friction has eventually discovered that some of that friction was load-bearing — the delay, the queue, the quarterly pause was where humans did their thinking. The reveal died at 4:32 PM because the queue it served had already been priced away. We optimized the wait out of the market. We are still finding out what the wait was for.

AUTHOR'S NOTE

This is speculative journalism written from an imagined 2036, examining where today's prediction markets could lead financial-market structure. The far-future specifics — continuous earnings pricing as standard practice, the retirement of the quarterly calendar, the 2031–2034 trading-desk scenes in the figures — are deliberate extrapolation, not forecast. The present-day foundation is real and sourced: the platforms' volumes and valuations, the Kalshi–CFTC litigation, the media deals, and the underlying theory (efficient-markets research on prediction markets, Soros on reflexivity, Perrow on tightly coupled systems) are all cited below. Where the line between sourced fact and projection runs, the citations mark it.