The trigger beat the loop.

In January 2024, Polymarket handled roughly four of every five dollars traded on prediction markets. By late summer 2026, Kalshi was handling roughly four of every five. Bloomberg ran the chart this week under a headline about Polymarket losing its lead, and the comment threads filled in the rest: network effects, regulatory moat, better execution.

Some of that is true. But a chart of market share is a picture of an outcome, not a mechanism. If you are a founder trying to learn something from it, the useful question is not who won. It is what turned the wheel, and whether that thing compounds or simply arrived.

So we ran it through the lens we use with clients: the business loop. Every real loop has four parts. A trigger that starts the behavior. A sequence of actions that creates value. A compounding asset the business accumulates. And retained value the customer carries into the next cycle. Miss one and you have a funnel, however circular the pitch deck looks.

That is a hypothesis, not a verdict. Desk research can see triggers, distribution deals, and the shape of a curve. It cannot see retention cohorts or what a trader actually feels. Where a claim below rests on public evidence we mark it Observed. Where it is reasoned from structure, Inferred. Where only customer research could settle it, Hypothesis.

01 · THE CURVE TESTRead the shape before the story

The first test in our diagnostic is the cheapest and most honest: look at the shape of the curve. A loop that compounds bends smoothly upward, because each cycle makes the next one easier. A business that grows when something outside it changes moves in steps.

Here is the same data, redrawn, with the outside events marked.

Look at where the share moves. It does not drift upward between events. It sits on a plateau, then jumps when something external happens: a distribution deal, a football season. The election spike in 2024 even falls straight back down once the occasion passes.

When Kalshi announced in September 2025 that it held 62.2% of global prediction-market volume, it noted the figure had been 3.1% a year earlier. That is a dramatic change. It is also exactly what a staircase looks like when you only compare the bottom step to the top one.

02 · THE TRIGGERTwo loops, two clocks

Strip away the branding and both companies sell the same thing: a contract that pays if something happens. What differs is the occasion that sends a person to open the app in the first place.

Polymarket's core occasion is the news. An election nears, a war escalates, a central bank meets. The trigger is real and powerful, and its prices became news in their own right during 2024. But the biggest occasions arrive on their own schedule, and between them demand thins out.

Kalshi's core occasion is the game schedule. By the fall of 2025, football accounted for more than 70% of Kalshi's trading volume, according to sports-betting analyst Dustin Gouker. Keyrock's 2026 comparison puts Kalshi's sports share near 85%, against roughly 39% for Polymarket, whose volume is spread across sports, politics and crypto. A game kicks off every week. The trigger never has to be manufactured.

Cadence matters more than most founders expect. A loop only compounds when it turns faster than the value it creates decays. A trader who places a bet every Sunday builds a habit. A trader who shows up once per election cycle has to be re-acquired every time. Even an average loop on a weekly clock will outrun an excellent one on a four-year clock.

Put the two side by side and the difference is not in the middle of the loop. It is at the front.

Polymarket's loop is arguably the more interesting one. A market that becomes the reference point for what is likely to happen has built something that compounds: the more it is cited, the more people trade it, the more accurate and citable it gets. That loop is real. It just turns at election speed, and in the United States it was locked out of the trigger entirely until Polymarket's return.

03 · THE STOCKVolume is not the thing that compounds

Here is where the chart quietly misleads. It measures notional volume: dollars changing hands. Sports, and parlays in particular, generate enormous turnover per person. A chart of turnover will favor whoever owns the highest-turnover occasion, whether or not their customer base is larger or more loyal.

Change the scoreboard and the race looks different.

The third bar is the one we would put in front of a skeptic. The World Cup was the largest prediction-market event on record, bigger than the 2024 election, and it was a trigger both companies could reach. Sacra estimates roughly $7 billion traded on Kalshi and $6 billion on Polymarket that June. When access to the occasion was equal, the gap nearly closed.

That is what you would expect if the advantage lives in trigger access rather than in a superior loop. It does not prove it. But it is the closest thing to a natural experiment the public record offers.

04 · THE EXPOSURESWhere the lead could leak

Our diagnostic runs three failure tests on every loop: can it keep compounding, can it keep getting fed, and does the company keep the benefit when it does. Applied to Kalshi's sports loop:

The first row deserves a moment. Renting distribution is often the right call early. It is how Kalshi reached tens of millions of retail accounts without building an audience. But a rented route to the trigger means someone else owns part of the environment your customer acts in. The moment their incentives diverge from yours, the loop keeps turning on their surface instead of yours.

05 · THE REVERSALWhat would prove us wrong

A strong claim should come with the evidence that would overturn it. Here are the two we would check first, and both are visible within months.

Neither is visible from outside with any precision. Kalshi can answer both in an afternoon.

06 · THE LESSONWhat a B2B founder should take from a betting war

Most founders we work with never think about prediction markets. They do think about why last quarter's growth came in a lump, and why the motion that worked in March stopped working in August. The pattern is the same, so are the questions.

1. What occasion sends a buyer to you, and how often does it fire?A budget cycle fires once a year. A failed audit fires once. A weekly report fires weekly. The cadence of your trigger sets the ceiling on how fast any loop can turn.
2. Who owns the surface where that occasion shows up?If a marketplace, an integration partner or a search engine routes the need to you, you are renting the front of your loop. That can be the right trade. It should be a deliberate one.
3. Is your curve bending, or climbing stairs?Plot new customers against the dates you signed a partner, launched a channel, or caught a seasonal wave. If every step has a date on it, you have growth. You do not yet have a loop.

None of this makes Kalshi's position weak. Owning access to a weekly trigger is a genuine advantage, and plenty of loops start life as a rented trigger before they learn to fire on their own. The point is to know which one you are running, because the investment that protects each is different.