Prediction markets are having a moment.
Users, VCs, regulators, founders. Everyone has an opinion. Most of them are optimistic. The narrative is familiar: better information, collective intelligence, democratized forecasting.
Some of that is true. Some of it is not new. And some of it is still very far from reality.
Not New. And Yet, Potentially Different
At a surface level, prediction markets are not a new invention.
We’ve had betting markets, derivatives, and various forms of speculation on outcomes for a long time. Wrapping that in tokens and smart contracts does not automatically make it innovative.
In many cases, what we’re seeing now is exactly that: existing products, slightly repackaged, with a blockchain layer on top (or beneath). Same mechanics, same incentives, same limitations.
The two most obvious differences from most existing products are the P2P nature and the ability to exit early by selling the token.
Both improve UX.
But are they fundamentally new? Or just features that could have been added by DraftKings or Robinhood?
But there is another angle, and this is where the real story is.
The idea that anyone can create a market, define an outcome, and allow others to take positions on it is meaningfully different. Removing or reducing the “house” as the sole market creator changes the structure of incentives. It opens the door to participation models that simply didn’t exist before.
So we’re left with two truths:
This is not new.
This could still be disruptive.
Both can be correct at the same time.
Two and a Half Models, Two Futures
Underneath the hype, there are three different business models emerging.
The first is centralized.
These platforms look and feel familiar. They operate similarly to traditional prediction or betting markets, with the addition of blockchain rails. Custody, matching, and market creation are still largely controlled.
The value proposition is incremental: better settlement, broader access, some efficiency gains.
There is nothing inherently wrong with this model. In fact, it does and will continue to work. It will likely dominate a large part of the market, precisely because it is easier to understand and regulate.
The second model is decentralized.
Here, the ambition is larger.
Anyone can create a market. Anyone can participate. The role of the operator is minimized or removed entirely. In theory, this is closer to the “pure” crypto narrative: permissionless, global, composable.
This is closer to the crypto narrative: permissionless, global, composable.
It is also where most of the real innovation could happen.
But it comes with obvious trade-offs. Without a central actor, questions around manipulation, insider trading, oracle design, and data integrity become significantly harder to solve. The system is more open, but also more fragile.
The third model is what most platforms actually are: hybrid.
They run a matching engine, they oversee operations, they’re the decision makers on a lot of what’s going on.
Again, nothing inherently wrong with this model. This often is the practical way to get off the ground with a roadmap to viable decentralization. But in other cases, this is the plan going forward – it’s easier than full decentralization, more practical, and keeps more of the value in the founders’ pockets (also, not inherently wrong).
Sometimes this is a transition phase.
Sometimes it is the business model.
Regulation Will Force a Choice
Prediction markets do not exist in a regulatory vacuum.
They sit at the intersection of gambling, derivatives, and financial markets. All heavily regulated, for good reason.
| There is a growing expectation that regulators will create bespoke frameworks for prediction markets, as they did for crypto. They won’t. At least not in any meaningful, paradigm-shifting way. We already have the frameworks. Prediction markets fit them squarely, we just need to place them properly within those frameworks. What very well may happen is minor adaptations to such existing regimes, to make them appropriate for on-chain settlement, more clearly define which activity falls within which basket (e.g. gambling vs derivatives), and MAYBE even a very pro-industry invention of a hybrid licensing regime allowing a platform to offer gambling products and derivatives under the same roof. But this is evolution, not reinvention. |
As the space matures, platforms will be forced to choose.
Centralized players will be forced to move toward full compliance or shut down.
Multiple licenses, jurisdictional structuring, full KYC, full control. and effective measures to identify, prevent and stop bad actors.
This path is complex, expensive and fragmented, but it provides legitimacy and access to larger pools of capital. We are already seeing an arms race between the largest crypto and TradFi players, building or buying platforms – those will be (either off the bat or over time) subject to very clear and robust rules of existing regulatory regimes.
Decentralized platforms will lean into permissionlessness. This aligns with the ethos of crypto, but also increases exposure to misuse, manipulation, and financial crime risks. It is harder to control, harder to supervise, and harder to integrate with the traditional system.
Both paths will coexist. But they will diverge more clearly over time.
The hybrid model will shrink – platforms will need to pick a side, or keep playing cat-and-mouse with regulators around the world.
Not because it’s wrong, but because it becomes harder to justify.
This Is Version 1.0
As we like to say in crypto – What we are seeing today is still “very early”.
The current wave is still heavily driven by attention, hype, and speculation. Much of the activity is closer to “trading narratives”, with institutions and data gatherers exploring the potential use cases and meaningfulness of information from markets.
Attention to the potential data value is already here, and reference to prediction markets in mainstream media is already being made, but ask yourself who can extract the “real data” -Strip out market makers, arbitrageurs, bad actors – to be able to distill the true convictions of the people.
That does not invalidate the space. It just puts it in context.
If anything, prediction markets today feel like a first iteration. A proof of concept showing demand, but not the final product.
The real opportunity is what gets built on top:
- hedging tools tied to real-world or onchain events
- insurance-like products with dynamic pricing based on market sentiment
- data layers that extract and aggregate probabilities as signals
- new forms of risk management embedded directly into DeFi
This is where prediction markets move from “interesting” to “useful”.
From “Fun” to Function
Today, a large part of the appeal is still the “fun” layer.
Elections, sports, headlines. It drives engagement, liquidity, and attention.
Over time, that layer will remain, but as a niche.
The real growth will come from serious use cases. Larger players, structured products, integration into broader financial and DeFi ecosystems.
That transition will not happen overnight. It will require sharpened product-market fits, better data, clearer rules, and learning from the V1.0 experience.
We’re not there yet, but this is where it starts becoming real.
………………………………………………………………………………………………………………
On April 15, 2026, I participated in a panel discussion titled “Can You Regulate the Future?” as part of “The Odds”, a Dedicated Prediction Markets Conference featured in Paris Blockchain Week.
In the panel my esteemed colleagues Ari Redbord, Mick Bransfield, Stephen Humenik, Andreas Glarner and I discussed the regulatory hurdles and opportunities facing prediction markets, and how they will adapt and help shape the future of prediction markets. Catch highlights from the panel, here.
Thanks to Tal Mor for organizing, Anne-Grace Kleczewski for the matchmaking, and Eylon Aviv for the spicy example on risk hedging, which I unabashedly 100% stole from him.
