“Markets don’t lie like people do. When people put money on the line, they stop lying. It’s amazing.”

– Tarek Mansour, CEO of Kalshi –

This line from Tarek Mansour, co-founder of Kalshi, is a strikingly confident one. It rests on the idea that markets can reveal something more objectively true than public conversation. But that assumption is worth questioning. After all, the history of finance gives us plenty of reasons to be cautious about the claim that money automatically produces truth.

Prediction markets have been, let’s just say, controversial for a while.

If you are not familiar with them, they are the gambling-like platforms you see advertised everywhere, which people insist are not actual gambling.

Jokes aside, prediction markets are platforms where people buy and sell contracts based on future events. These can be almost anything these days, ranging from things like whether the US will confirm aliens exist, election results, sports outcomes, or what will happen before the release of GTA VI, all the way to more controversial topics. The market price is usually treated as a rough estimate of how likely that event is to happen. 

The most standard payoff structure is binary contracts. If the event happens, it pays out, if it does not, it expires worthless. Traders can buy and sell at any point before the outcome is known. Because of this, the market price is often read as a rough estimate of how likely the event is to occur. 

A contract trading at 65 cents, for example, is usually interpreted as implying a 65% chance for the events to happen. This is why prediction markets are considered to be a reflection of collective belief.

So how exactly is this not considered gambling, one (many, really) might ask, and can we trust these markets to predict the future? 

Gambling, Finance, or Something in Between?

While Arcanum Ventures was researching for this piece in Europe, one issue immediately got in the way. Prediction market platforms had been blocked by the country’s regulators over alleged illegal gambling. This means that access is effectively cut off unless the user is outside the country or can bypass the geoblock.

Whether prediction markets count as gambling at all is still a deeply contested question around the world, and the classification of these platforms varies country by country. Supporters argue that they are closer to financial markets. Critics see them as gambling in a more legally protected form.

While opinions remain divided, these platforms have managed to persist exactly because their legal status remains so ambiguous. This ambiguity is partly rooted in their history, as discussed later on, and has often allowed them to avoid being clearly classified as gambling.

Firstly, their operators do not present them as bets at all. Instead, they frame them as “financial contracts” whose value depends on the outcome of a future event. Which might basically sound like gambling, but the argument is that this is actually a regulated market product.

Because of this framing, they often fall into different legal categories than conventional gambling. Gambling is usually understood as risking money on an uncertain outcome for a payout, while derivatives law treats certain event-based contracts as financial instruments that can be traded, priced, and, in some cases, used for hedging.

To an ordinary user, the experience may still look a lot like betting. But in regulatory terms, the difference often comes down to classification, whether it is a wager or a financial contract. 

Global Markets vs Local Laws

Because of this ambiguity, there is no single stance on prediction markets globally. Whether they are accepted or not begins with how the country classifies them.

In places where regulators accept the idea that these platforms are offering financial contracts tied to future events, they are more likely to be treated as a form of derivatives market. That is largely the direction the United States has taken, where event contracts are often framed as swaps or other regulated market products.

In much of Europe, the approach has been a lot more skeptical. There, consumer-facing prediction markets are more likely to be seen as unauthorized gambling, especially when they operate outside local licensing systems.

Asia is less uniform, but the general tendency is stricter. In much of the region, prediction markets are more likely to run into gambling laws. Hong Kong has recently stated that betting on sports through prediction market platforms may be illegal, which reflects a broader regional instinct to treat these platforms with suspicion rather than regulatory optimism. Elsewhere, they often remain in a legal gray zone, neither fully accepted nor clearly prohibited. 

But if there is so much pushback around the world, how did prediction markets become this large? The answer is that the story did not begin with Kalshi or Polymarket.

From Academic Experiment to Market Industry

Humans have always been drawn to betting, and informal wagering on elections and wars has been part of human history for centuries, reaching back at least to sixteenth-century Italy.

But the prediction markets we know of today were not intended to be a commercial betting empire. They began as a small academic experiment in forecasting in 1988, with the launch of the Iowa Electronic Markets at the University of Iowa. Created by economists Robert Forsythe, Forrest Nelson, and George Neumann, the project was designed to test whether markets could forecast political outcomes better than polls.

They themselves were surprised by just how accurate the first results were. As the experiment continued, the researchers found that the Iowa Electronic Markets were closer to the final outcome than polls nearly three-quarters of the time, especially well before election day.

Seeing the value in this, the CFTC decided to allow prediction markets to operate as a nonprofit research and education project through a no-action letter in the early 1990s, which gave the model legitimacy in the United States.

From there, prediction markets gradually evolved from academic forecasting tools into online platforms of consumer-facing markets, dedicated for trading on public events.

Today, that process has effectively produced two giants: Kalshi and Polymarket. However, their founders had different motivations for their launch than just simple academic curiosity.

The Founders Who Wanted to Price Everything

Kalshi began with a frustration that had been building for years inside traditional finance. Tarek Mansour and Luana Lopes Lara (who is now the world’s youngest self-made female billionaire at 29) had seen firsthand how forecasting tools on Wall Street were treated like privileged instruments. These powerful tools locked behind expensive gates were accessible only to a narrow circle, and people working on understanding risk were often excluded from this circle.

This dissatisfaction became the spark to their idea. In 2018, Mansour and Lopes Lara set out to build a platform that could make forecasting more open and democratic. For Mansour, the ambition was bigger than finance alone. He openly shares his views on wanting to expand the idea to become “as big as possible”, since he firmly believes that markets reveal conviction more honestly than an ordinary debate ever can, as they force people to put money behind what they believe.

“ The long-term vision is to financialize everything and create a tradable asset out of any difference in opinion.”

– Tarek Mansour, CEO of Kalshi –

For Polymarket, the motivation was similar, but taking the same core prediction-market idea and making it faster and more internet-friendly through crypto rails. Shayne Coplan (who also briefly held the title of world’s youngest self-made male billionaire title at 27 in 2025) founded Polymarket in 2020, explicitly positioning it as a way to bring users closer to major events and to turn public attention into tradable activity.

The two giants have been growing ever since. Polymarket and Kalshi together accounted for roughly 90 percent of a record 2.35 billion dollars in weekly trading volume in late October 2025. As of 2026, Kalshi is valued at roughly $22 billion, while Polymarket is in fundraising talks that could value it at around $15 billion.

Two takeaways here: 

1) Leaning into a trend that seems a little absurd might somehow turn you into the world’s youngest billionaire, and

2) It seems that founders claim that the biggest argument for prediction markets is their ability to provide a transparent and supposedly truthful platform for pricing uncertainty and aggregating collective expectations about the future.

How Accurate Are Prediction Markets?

Let’s go back to Mansour’s quote, which argues that once people have money on the line, they start revealing what they really believe.

This, in theory, should make prediction markets work well. Take a market price and constantly update it through buying and selling, and you should get fairly transparent data of the public opinion. There is definitely evidence for this idea. Under the right conditions, prediction markets can actually forecast outcomes surprisingly closely. 

For example, at the start of the Iraq War in March 2003, traders were about 90% sure he would be out by April and 95% sure by May or June. He was in fact deposed in April 2003, so the market was directionally and temporally very close. Or during the 2026 Oscars, reporting after the ceremony found that Kalshi traders got 18 of 24 categories right and Polymarket got 19 of 24 right when measured by the favorite at the end of Academy voting. Just recently, both platforms correctly predicted the outcome of Hungarian elections as well. 

However, prediction markets are not infallible. During the Brexit referendum, for instance, markets largely favored Remain, but the final result was a vote to Leave. A similar miss occurred in the 2016 U.S. presidential election, when prediction markets broadly pointed to a Clinton victory before Trump’s win.

But if we are to follow Mansour’s views, then how come in some cases the predictions fail? 

Well, to put it simply, it is because we are no longer inside a curated academic setting anymore. 

At their best, prediction markets work because they aggregate scattered information. Ideally, each trader brings some fragment of knowledge and interpretation. It depends on the market having enough participation and enough diversity of information to keep the price honest.

Thin Markets

A major issue is when trading is thin and the liquidity is weak. In a deep market, a mistaken price tends to get corrected because someone has both the incentive and the ability to trade against it. In a thin market, the prices can be moved too easily.

Echo Chambers

A second issue is when information is correlated rather than independent. Prediction markets are often described as “the wisdom of crowds,” but that wisdom only works when the crowd is not simply echoing itself. In that case, the market is just re-pricing the shared narrative, and becomes vulnerable to herding.

Complex Situations

Another problem is that some events are simply hard to price cleanly, even for a functioning market. Prediction markets do best when the event is more near-term and can be resolved by a straightforward fact. But in reality, there might be a hidden turnout, or last-minute shifts, or voter groups that are hard to observe properly. I.e., it is easy to miss events unfolding in the background.

And this is mostly to do with the market mechanics and the collective human nature. The truth is that not every trader has the same amount of information, or wants to be an honest player in the game.

Money Does Not Automatically Mean Truth

Money on the line creates an incentive pressure. And incentive pressure can produce very different behaviors in different individuals. Incentive pressure does not automatically produce honesty. For some participants, it produces the opposite: strategic manipulation, insider positioning, and deliberate distortion of the information the market is supposed to reveal.

History offers plenty of reasons to be cautious. The broader financial world is full of examples where high stakes did not eliminate deception. WorldCom’s accounting fraud, pump-and-dump schemes, and countless cases where profit created strong incentives to distort reality.

Prediction markets are not immune. There has recently been a lot of scrutiny around Iran-related Polymarket bets, where unusually well-timed positions raised concerns about insider-linked advantage.

Profiting From Crisis 

The suspicion of insider trading was not the only source of scrutiny. The kinds of events people could bet on also raised deeper concerns. It is one thing to trade on election outcomes or interest rates, and it is another to turn deaths, coups, disasters, or political collapse into tradable markets.

For many, the problem is moral as much as financial. Public suffering and geopolitical instability become opportunities for profit, while every uncertain event, no matter how grim, is reduced to a live number people can refresh and trade on for gain.

The discomfort is only heightened by the fact that regulation often moves more slowly than public opinion. Enforcement tends to be reactive rather than preventive. And once a platform is framed in the language of finance and backed by institutions powerful enough to defend it, it can acquire legitimacy faster than society is able to reckon with its consequences.

In fact, recent news shows that some regulators have actively tried to stop these platforms altogether. Arizona filed criminal charges against Kalshi in March 2026, accusing it of operating an illegal gambling business, while other states, such as New Jersey, Massachusetts, and Nevada, have also moved to block or restrict parts of its business. Yet even those efforts have not forced the industry into retreat.

It is also worth noting that Donald Trump Jr. advises and invests in Kalshi and Polymarket, and that the Trump administration has supported prediction markets in the broader federal-state dispute. This may help explain why the industry has not simply folded under regulatory pressure.

The Case for the Technology: Infrastructure

There is also a more constructive version of this story.

The technology may feel dystopian in some areas, but the underlying infrastructure can be used for something genuinely valuable. These systems can offer a way to aggregate dispersed information quickly and help people make decisions before outcomes are known.

Researchers have already explored similar possibilities. Work on epidemic prediction markets, for example, has examined how real-time forecasting systems might help anticipate the spread of infectious disease. Adjacent forecasting efforts in food security and anticipatory humanitarian action likewise show how early-warning systems can support better preparation before a crisis escalates.

There is also a technical reason prediction markets could be useful. Platforms like Kalshi and Polymarket now expose useful information, such as APIs, market data feeds, order books, and execution systems, that allow other developers to build on top of them.

Kalshi’s developer documentation, for instance, presents the platform as an exchange API for real-time market data and trade execution, with support for REST, WebSocket, FIX, and SDKs. Polymarket’s documentation does something similar, offering APIs, SDKs, live data access, and developer tools for building directly on its market layer. In addition, its public GitHub repositories also reveal exchange contracts, CLOB clients, and adapters for oracle-based resolution.

That means this technology is increasingly becoming a reusable infrastructure layer for several kinds of products.

Trading interfaces and broker integrations

Because these platforms expose APIs and execution rails, other applications can build custom front ends, embedded event-trading modules, or broker-style access layers on top of them.

Real-time data products

Since both platforms expose market data and order books, developers can create dashboards, analytics tools, monitoring systems, and research products that track probabilities, momentum, liquidity, and event sentiment in real time.

Protocol-level prediction infrastructure

Newer projects are trying to move beyond the standalone-platform model and turn prediction markets into shared backend infrastructure. The idea is that multiple applications could plug into the same prediction-market layer, rather than each rebuilding the full stack from scratch.

Oracle and resolution systems

One of the hardest technical problems in this space is determining what actually happened in the real world. Polymarket’s architecture shows explicit oracle-resolution infrastructure, including adapters tied to external resolution systems. In other words, part of what is being built here is machinery for translating messy real-world events into machine-settleable outcomes.

Seen in that light, the positive case for the technology is maybe not the market itself, but the infrastructure that can be used as a foundation to be built on.

The Bet Going Forward

So, are prediction markets true and honest? Not exactly. They do not escape the same incentives, blind spots, and power dynamics that shape every other market. But they can still reveal something real, and sometimes useful.

Looking ahead, prediction markets are unlikely to disappear. If anything, they are becoming more embedded, both as platforms and as infrastructure. The more important question is what role they end up playing. They could remain speculative venues, or they could evolve into tools that support decision-making in critical areas. Most likely, they will become some combination of both.

One thing is certain, Arcanum Ventures will be watching the space closely to report on what comes next. That is one bet worth placing.

Continue the Conversation

Arcanum Ventures stays close to the ideas, people, and systems shaping tech before they become mainstream consensus. From privacy tech and AI to data infrastructure and token design, we work with founders and teams building in complex, high-stakes environments.

We track these questions in public through podcasts, videos, and original writing, while working directly with the people building inside them. Join us as a speaker, tune in to the conversation, or keep pace with what’s moving across tech. 

If you are building something difficult and want a second set of experienced eyes, we would love to hear from you.