Interview with Johann Evrard, CEO, Predictive Labs.

Start with the word... 'Insider' comes from the stock market, where it means something precise: you hold information the public does not, and you have a duty not to trade on it. A prediction market runs the other way: it exists to pull what people know into a price, and it does that by paying whoever knows more than the crowd. So an edge is the fuel the market runs on, not a flaw in it. You read the filings, you talk to people in the field, you spot a situation the market has not caught up to, and you get paid for being early. That is the whole point of it!

Which is what makes 'insider' so slippery here.

Take the clean case: a Google engineer charged for trading on Polymarket off confidential internal Google data, the kind the market was never built to price in. The guy on the other side, the one who put $50 down off a headline, is the one who pays. So far, easy.

Now take the cases the policy is really aimed at, the geopolitical and law-enforcement ones. Some are just as clear, only worse: a soldier who bet off classified intelligence about the Maduro capture was stealing information he had a duty to protect, full stop. But most are not that clean. Between outright theft and the sharp analyst reading open sources sits a wide middle: someone who simply understood a situation better than the crowd, which is the exact thing the market is built to reward. From the outside, an honest informed trade and a leak often look the same, so most of the time you cannot say which is which. And the market needs that middle to stay full, because that is where the information comes from: police it too hard and you scare off the legitimate trader along with the crook. So the real question under the policy is which kind of knowing counts as off-limits, and a screen has to settle that before it can screen anything at all...

Employer disclosure is the tool on the table. But what does it actually screen? It asks where you work, and Kalshi has been clear the answer is self-reported, not verified unless an investigation is warranted. So walk it through: say your edge comes from your job ; you are the Google employee kept off a Google contract, and the form catches you, at least if you fill it in honestly, because the link is sitting out in the open. Now say it comes from somewhere else, for instance a tip from a contact, or a read nobody can trace back to an employer. You fill the form in truthfully, you trade, and nothing flags. The sophisticated one does the same and just routes around it. So is it sufficient? No, the trader who actually matters is the one it never sees. But it is not nothing either: it deters the casual case, it lets Kalshi block an obvious conflict before the trade, and a false answer turns a shady trade into a documented lie an investigator can use later. Call it a floor, not the whole building. It works at the obvious end of the spectrum, and stops there.

There is a 2nd place to look, one that complements the form rather than replacing it. Disclosure asks who you are ; but a real edge usually surfaces in the trading itself, before anyone knows the cause ; a sudden spike in volume, or a whale, a single position big enough to move the price by itself. At Predictive Labs our job is to surface those moves, the ones that break hard from a market's own baseline, across every venue instead of one. It will not tell you why, informed money, a leak, or a random blip, that part is your call ; and it will miss the careful insider who keeps his position small on purpose. What it does is show where to look, which the form on its own cannot...

Either way, the deeper issue is that a disclosure rule lives on one venue, and the behaviour it targets lives across all of them. Put friction on the most sensitive markets at the most regulated venue, and a share of that flow just goes where the friction is lower, including crypto-native venues where people trade under a pseudonym and an employer form means nothing at all. Tellingly, the cases actually charged so far, the Google engineer and the soldier, were already on that pseudonymous venue, not on Kalshi. So at least a share of the flow moves rather than stops, and the tighter one venue polices, the more it pushes toward the ones that do not. Whether that trade is worth making is a regulator's call...

So, yes, there is a real near-term limit, and it is worth being straight about it: in a lot of the world, outside the US and Europe, a stablecoin is the main way most people can put money into one of these markets. Where the rails have not spread, people cannot get in, and growth stalls. But that limit is about access, how far the rails have physically reached, not the size of the pool behind them.

The size of the pool is a different question, and that is where the $300 billion against total banking assets gets read the wrong way. What is that comparison really measuring? It sets one speculative instrument against the whole global banking system, so of course it comes out under 1%! But that is the wrong test. The real question is whether there is enough liquid money at the on-ramp to clear the positions people actually want to take, and there plainly is: all the regulated US venues together did on the order of $25 billion of volume last year, a sliver of the stablecoins already sitting on-chain. For me, stablecoins are just the on-ramp, not the ceiling.

So if the cap is not the size of the pool, where is it? In three places. It is regulation, whether a venue may offer a market in a country, and whether someone there may legally take part. It is distribution, everything between an ordinary guy and their first position: whether the rails even reach them, then the sign-up, the ID checks, funding a wallet, learning the interface... And it is trust, whether people believe the market will resolve honestly and pay out: you either trust a venue with your money or you walk! Regulation, distribution, trust. In my opinion, that is what holds growth back, while the stablecoin pool sits there with room to spare.

On who ends up dominating, I think the premise of one simply beating the other is, in a way, a little off, because for now they are mostly serving different people... The pseudonymous trader putting money on a global event from a place with no regulated venue is rarely the same customer as the institution that needs a licence and a fiat settlement account before it can touch anything.

For a while they grow on separate tracks, and then the tracks start to blur. Kalshi is fully CFTC-regulated and now takes stablecoin deposits in USDC. Polymarket went the other way and bought a licensed US exchange to come back into the country as a regulated venue. Regulated venues reaching for crypto rails, crypto venues reaching for a licence... So if you push me, I lean hybrid, meaning the model that wins is a blend, not one camp beating the other.

How fast it moves is mostly a regulatory question, and regulation does not only travel one way. For instance, the opening prediction markets are enjoying in the US right now is tied to the current administration, and it is still contested by the states, so it could very well tighten again under the next one. So you are likely to see it happen in patches, one market at a time, some of them sliding backwards for a stretch, rather than a clean line up.