Of everything crypto has produced, prediction markets may be the application that has traveled furthest outside crypto itself. The cynical reading writes itself: the technology's breakout consumer product is a betting parlor. That reading is not wrong. It is incomplete.
Start with what the label "crypto casino" obscures. On Polymarket, what actually lives on-chain is the money and the positions: the collateral, ownership of outcomes as conditional tokens, and settlement once a market resolves. Traders sign their orders cryptographically, and the platform's own documentation is explicit that the operator cannot execute trades a user has not authorized. The order book and the matching engine, by contrast, run off-chain and are centrally operated. It is a hybrid design: centralized speed where speed matters, on-chain custody where trust matters.
The target customer is, plainly, the bettor. But let the general public bet on nearly everything and something else emerges inside the casino: a mechanism essential to all economic activity — the transfer of risk. In July the New York Post told the story of Jason and James Jiang, brothers who run 28wishes, an ice-cream shop in downtown Los Angeles. When the temperature slips below 70°F — about 21°C — they say sales drop roughly 20%. So since April they have staked about $20 a day — some $600 a month — on cool-weather contracts on Kalshi, Polymarket's centralized, regulated American counterpart. By their own account the positions have returned up to $1,500 in a month, with a best day of $800, against monthly rent of $3,500 — about 43% of the rent, as the headline put it. Those figures are the shop's statements, not audited results. The structure is what matters: the brothers assembled an insurance policy without an insurer.
Insurance, reduced to its economics, is a transfer of risk. Traditionally that takes a capitalized insurer, pooled premiums, underwriting, an adjuster and a claims process. On a prediction market, much of that chain disappears: the counterparty is a speculator or a market maker, and settlement is binary and standardized. The speculator ends up playing the economic role of the insurer — worth pausing on, because speculation is usually cast as the villain. Yet without someone opposite willing to carry the risk you want to shed, there is no hedge at all. In principle, the more participants a market draws, the closer its price sits to a fair probability — and that is what makes hedges available here that exist almost nowhere else. No adviser offers a contract that pays out if a given candidate wins an election.
The limit is just as instructive, and it is why these markets worry regulators. Classical insurance demands an insurable interest: you may only insure what its loss would genuinely cost you. You cannot insure your neighbor's house — a rule that exists precisely so that a contract cannot create an incentive to cause the loss. Prediction markets impose no such condition. The economic line between hedging and betting collapses into one question — does the trader actually carry the underlying risk? — and the regulatory problem sharpens where a trader can influence the outcome, trade on inside information, or reach the market's resolution process. An ice-cream shop does not control the weather. Not every participant in every market is so safely powerless.
None of this is conceptually new. Weather-exposed companies have had exchange-listed weather derivatives at CME since September 1999. What changed is not the instrument but the ticket in: an app, a few dollars a day, no desk, no broker. Nor is hedging the main event on these platforms — sports has made up about 80% of Kalshi's trading volume since July 2024, by Pew Research Center's count, and open disputes over insiders and market resolution have followed the sector. The ice-cream shop is an anecdote. It is also a demonstration of what this infrastructure makes possible.
Meanwhile, the infrastructure is being institutionalized. Intercontinental Exchange, owner of the New York Stock Exchange, announced a strategic investment of up to $2 billion in Polymarket in October 2025; in February 2026 it launched Polymarket Signals and Sentiment, becoming the exclusive provider of Polymarket's data to institutional capital markets — crowd-sourced probabilities, normalized and delivered alongside traditional market feeds. CME said on February 13 that 100 million of its event contracts had traded since their December debut, eight weeks earlier, reaching retail through brokers that include FanDuel's and DraftKings' new prediction apps. Step back and a loop appears: the casino produces a probability, and institutional finance buys it as data. Bettors are financing, mostly unknowingly, a public good — a live estimate of the odds.
There is a reading of Bitcoin that belongs at the end of this story. Every hedge above depends on someone: a counterparty who must stay solvent, an oracle or resolution source that must rule, an institution that must pay the claim. Bitcoin held for the long term is often described as insurance at the scale of a person or a state — yet it indemnifies no one. There is no claim to file, no resolution to vote on, no counterparty to keep honest. The hedge is the holding itself. This is general market information, not a recommendation to buy or sell any asset.