Something strange happened to the American market structure in the spring of 2026. Kalshi, a prediction market, became the first company in U.S. history to offer CFTC regulated perpetual futures. Polymarket, another prediction market, launched 10x leveraged perps on Bitcoin, Nvidia, and gold within weeks. Meanwhile, Robinhood, Coinbase, and Kraken (three companies that began as a stock brokerage and two crypto exchanges) all added prediction markets. Offshore perpetuals had already grown from $28 trillion of annual volume in 2023 to more than $90 trillion in 2025, becoming one of the fastest growing financial products ever created. Now, the contracts are escaping crypto entirely, extending to tokenized equities, commodities, and FX, trading around the clock while the NYSE sleeps.
Squint at that list and the pattern resolves: every retail venue in finance, whatever it started as, is converging on the same product suite: leverage, binary outcomes, continuous access, no expiry. Convergence like that is never about the products. It's about the customer. Every one of these firms has concluded it is serving the same person, and that this person wants one thing in escalating doses: convexity, small stakes with a shot at life changing outcomes.
The conventional explanation is the bull market, and it's not wrong. Risk appetite has always been procyclical: rallies mint confidence, confidence mints leverage, and equities spent the past year clearing targets. But the procyclical story has a problem: it's been true forever, and it doesn't explain why this cycle produced a permanent new product architecture rather than the usual burst of margin debt.
Something else is running underneath the cycle. My view is that a growing number of people have concluded that income can no longer buy the life they were promised, and that saving cannot close the gap, so risk has become the only rational instrument left.
The Arithmetic of Falling Behind
Since 2020, U.S. asset prices have run far ahead of wages: equities up on the order of 90%, home prices up roughly half, while nominal pay rose a fraction of that. The house price to income ratio sits near record highs, roughly five to six times income, against a long run norm closer to three and a half. The median first time homebuyer is now around 40 years old, a number that would have read as a misprint a generation ago. Marriage, children, and retirement; the milestones that define "keeping up," have all drifted later, and surveys of younger cohorts consistently find majorities who believe the traditional path (salary, savings, compounding) simply cannot get them there.
Now run the personal math that millions of people run implicitly. If the down payment you need is $150,000 and your maximum saving capacity is $500 a month, saving is not a slow path to the goal; it is no path: twenty five years, during which the target itself inflates away from you. Against that arithmetic, a portfolio of convex bets is the only strategy whose payoff distribution includes the goal at all.
Behavioral economics predicted this half a century ago. Prospect theory's central result is that people are risk averse when protecting gains and risk seeking for small losses, with asymmetric payoffs, and that "loss" is measured relative to a reference point. If your reference point is the life your parents had at your age, the home they owned, the single income household, the retirement date, then a large share of two generations is operating permanently in the loss domain. Prospect theory says people in the loss domain gamble to get back to even. The research on relative deprivation says the same thing from another angle: perceived inequality and unattainable benchmarks reliably increase risk taking. The lottery literature has documented for decades that ticket purchases are concentrated among those for whom conventional wealth building is foreclosed.
None of this is new behavior. What's new is the size of the population the arithmetic has pushed into the loss domain and the quality of the instruments waiting for them. Which brings us back to the product explosion and the causality question that determines whether this thesis holds: Did the products create the demand, or did they find it?
Products as Convexity Vending Machines
Much of this wave was unlocked by regulators: sports betting needed a Supreme Court decision, prediction markets needed CFTC approval, and U.S. perps needed a designated contract market. But legalization only explains timing. It doesn't explain velocity: why sports betting scaled to a hundred billion dollar handle within a few years, why event contract volume went vertical, or why perps became a $90 trillion market offshore before any U.S. regulator touched them.
Latent demand at that scale isn’t manufactured by product managers so much as it’s surfaced by them. The instruments map one to one onto the psychology:
Perps deliver maximum exposure per dollar of savings constrained capital.
Prediction markets deliver binary convexity wrapped in the dignity of "having a view."
Zero day options deliver the same on the S&P.
Memecoins deliver it with a community attached.
These are convexity vending machines, each calibrated to a customer whose defining constraint is that their stake is small and their target far. And the past two quarters have produced the natural experiment that distinguishes this from a bull market artifact. Crypto crashed total market cap down 20% in Q1, spot exchange volume dropped nearly 40%, and the memecoin economy is dead. If risk culture were merely rally reflexivity, speculative volume should have died with it.
Instead, it migrated: prediction markets grew, perp venues expanded into equities and commodities, with open interest rising every quarter, and brokerages raced to add event contracts. The risk appetite survived the destruction of its favorite asset class by changing instruments, which is exactly what a structural driver does and a cyclical one doesn't.
What Would Falsify This
The thesis rests on three conditions, and each one requires testing.
The first concerns persistence. If speculation is driven by a broken affordability ladder rather than by the cycle, the relevant measures (speculative participation per capita, perpetual futures open interest, event contract volume, and options share of retail flow) should keep rising through the next full equity drawdown, not only through a crypto-specific one. If they collapse alongside the S&P, the procyclical explanation wins.
The second concerns cross-section. Speculation intensity should be highest in the cohorts and countries where the milestone arithmetic is worst: among the young, the urban, and the priced out, and it should be lower where housing costs and wages still roughly balance. Early evidence points in that direction, though the prediction deserves a proper test. If intensity turns out to track income or financial literacy more closely than affordability, the mechanism is wrong.
The third concerns reversal. If affordability genuinely improves, with rates falling, housing supply expanding, and wage growth outrunning asset returns, the cohorts that regain access to the conventional ladder should visibly de-risk. If they keep gambling anyway, the culture has outgrown its cause, and that is a different (and darker) essay.
The Uncomfortable Conclusion
Dismissing this shift as a moral failing or cynical predation ignores the structural reality at play. For a generation priced out of traditional wealth accumulation, the 'casino' is a rational, high-stakes adaptation to a strategy that arithmetic has rendered obsolete.
We are witnessing a permanent restructuring of the financial landscape: every brokerage, exchange, and app is converging into a single, seamless machine for tradable conviction. These platforms did not manufacture this demand; they merely surfaced it, underwritten by years of asset appreciation that outpaced the paycheck. The venues are merely downstream; the gap is the product.
