Esports Betting in America: Seven Years of Saying “Not Yet” and a Product Still Looking for Its Users

A Packed Arena and an Empty Order Book

Seth Young tells a story he uses to open every conversation about his market. A sold-out arena for a League of Legends match. Chants rolling down from the stands toward the stage. Several thousand people shouting one player’s name in the thirtieth minute of a game. Then he points at the rest of the picture: the trading volume that prediction markets recorded for that same match.

It is so small it is almost invisible.

Young is the chief executive of ROLR. He played Counter-Strike 2 competitively before moving into operations, then founded High Roller, the predecessor product he ran in markets outside the United States. He says he first declared that “the US esports market is not there yet” seven years ago. Seven years later, he is saying exactly the same thing.

In an industry where every speaker wants to be the first to announce a new wave, a chief executive repeating a pessimistic sentence for seven years is a more notable fact than any growth chart.

ROLR Is Not a Sportsbook

The first distinction matters: ROLR operates in prediction markets, not fixed-odds sportsbooks. Users in prediction markets buy and sell event contracts — who wins, who loses, total maps, match duration — at prices that move with supply and demand. Users of fixed-odds sportsbooks place a bet at a price set by the bookmaker, who carries the risk on the other side.

That difference sounds terminological, but it determines the entire business model. A sportsbook needs capital to balance its book, limits for individual customers, a pricing team, and licences in each state with a different filing file. A prediction market needs liquidity, and liquidity cannot be bought with advertising.

Young positions ROLR between three groups of competitors. DraftKings, FanDuel and Fanatics are traditional sportsbooks with enormous infrastructure. Kalshi is an event-contract platform regulated at the federal level. ROLR chooses not to fight the first group head-on, and does not try to imitate the second.

His phrasing is telling: ROLR does not aim to take the entire pie, only to claim its “fair share”. That is the language of someone who has watched the money-burning wars of American sports betting and concluded that the winner is not the biggest spender.

The Regulatory Map and the Gap in the Middle

The US regulatory picture after the Professional and Amateur Sports Protection Act was struck down in 2026 is a maze. Each state decides whether to allow sports betting, to what degree, and for which product types. Kalshi operates under the Commodity Futures Trading Commission. DraftKings and FanDuel operate under state gaming commissions.

ROLR sits in the border zone. That zone has the advantage of being ignored by the giants, but the disadvantage that whenever a regulator reclassifies a product, an entire catalogue can be affected in one afternoon.

Young does not address this risk directly. But he says something indirect that I consider more important: legal friction limits available products, available products limit liquidity, liquidity limits user experience, and poor user experience limits growth. It is a closed causal chain, and cutting it at the last link is the wrong approach.

Three Frictions That Keep Viewers From Betting

The gap between US esports viewership and US esports betting volume is one of the widest in the entire sports ecosystem. From years of following competitions, I see it explained by five frictions.

The first is fragmented scheduling. A tennis tour has a fixed annual calendar; a basketball season is published a month in advance. Esports schedules shift constantly, tournaments overlap, teams compete across time zones, and a match sometimes starts two hours later than announced. For someone who wants to place a bet, that is operational risk, not sporting risk.

The second is data. A prediction market needs real-time data feeds clean enough to settle contracts. In esports, the data sits with the publisher. Whether third parties can access it, how fast, and whether they are licensed to use it, is a negotiation question rather than a technology question.

The third is fan structure. Esports viewers are young, used to free platforms, used to account sharing, and — most importantly — used to following players as characters rather than following teams. Bets driven by personal emotion are harder to price than bets driven by team analysis.

The fourth is integrity culture. A segment of the esports audience views betting with suspicion, because memories of match-fixing during the skin-betting era remain raw. Trust is not built with licences; it is built with time.

The fifth is the advertising channel. US esports teams and tournament organisers have not developed the habit of weaving financial products into content. They sell jerseys, fan packages and broadcast rights. They do not yet sell audience attention the way European football clubs have for decades.

These five frictions do not exclude one another. Together they explain how a sold-out arena can sit next to a near-empty order book.

The Liquidity Paradox

This is the most technical part of the story and the least discussed.

A prediction market is only useful when its prices reflect probabilities. Prices reflect probabilities when enough participants trade on both sides. Enough participants appear when a product has regular users. Regular users appear when the product is good. The product is good when liquidity is deep enough. That circle has no obvious starting point.

Financial exchanges solve this with market makers, who quote both sides at a spread in exchange for risk. Market makers need capital, risk models and hedging capacity. In esports prediction markets, risk models are far harder to build than for equities, because a young roster can change completely after one transfer window, and because there is no long, clean historical data series to train on.

In other words, ROLR is not only competing on product. It is competing on its ability to conjure a market out of nothing. That is a problem even large financial exchanges take years to solve.

The Unit Economics Problem

Here Young makes a specific, checkable claim: ROLR spends in a measured way, and the measurement shows positive return on ad spend.

More precisely, ROLR works with Spike Up Media, a multi-vertical lead generation firm. Spike Up Media is also a major shareholder in ROLR. Over five years running High Roller in markets Young himself describes as weaker than the United States, the pair recorded positive return on ad spend.

Reading that carefully reveals three layers.

First, the relationship between ROLR and Spike Up Media is not an ordinary vendor–client relationship. When a major shareholder is also an acquisition partner, both parties sit on the same incentive line. That lowers negotiation costs and partner-churn risk, but it also lowers the objectivity of the performance data.

Second, five years of positive return is long enough to strip out short-term luck, but not long enough to prove scalability.

Third, precisely because those markets are weaker, user acquisition there is cheap. A weak market has few competitors bidding up ad prices, few substitutes for users, and low product expectations. The same budget buys more users. But those same users also carry lower lifetime value, because disposable income and engagement are lower.

Young says he spends “surgically”, and the phrase fits the model: targeting segments where acquisition cost sits below lifetime value, rather than buying growth at any price.

The Big Pie and the Patient Share

Young refers to a large and growing pie and stresses that ROLR only needs its fair share. It is a plausible argument, and also the one I want to test hardest.

When a company says the market is large enough that a small slice will do, there are usually three possibilities underneath. First, the company genuinely believes in the market size and chooses sustainable growth over fast growth. Second, the company lacks the resources to fight for a bigger slice, and calling it a strategic choice is polite phrasing. Third, the company has not yet found a way to take a bigger slice.

There is not enough evidence to place ROLR firmly in any category. One detail tilts toward the first: five years of positive return in weaker markets is evidence of spending discipline, not of limited ambition.

From years of watching mid-tier esports teams across several regions, I see a striking parallel with the structure of small tournaments. Where there is no large rights money, organisers survive by optimising operating costs and squeezing value from every small sponsor. They do not try to take the big events’ share. They build a micro-ecosystem robust enough to survive season after season. That is the kind of discipline an immature market rewards rather than punishes.

The “Market Isn’t There Yet” Trap

Here I have to say what a conventional piece would skip.

The “market isn’t there yet” argument sounds mature. It creates the impression of a seasoned operator who refuses the hype and knows how to wait. But it can also be a cognitive trap, because it places all responsibility on the context rather than the product.

Three possibilities are hidden by that argument.

First: the market has arrived, but the product is wrong. If US esports viewers are not betting, it may be because they do not want to bet through the current product structure — too many steps, too much financial jargon, too much friction in deposits and withdrawals — rather than because they are not psychologically ready.

Second: the problem is event integrity. A prediction market works when event outcomes cannot be influenced by market participants. In esports, match-fixing cases at academy and regional level have been documented in recent years. If users do not trust the integrity of the match, no return on ad spend can save the product.

Third: real-time data is not clean enough. Contract settlement depends on data. Esports data depends on publishers. Publishers can change APIs, change terms of use, or change data-sharing policy season by season. A prediction market built on unstable infrastructure carries foundational risk, and users do not forgive foundational risk.

If any of those three is true, then “the market isn’t there yet” is another way of saying “the model isn’t right yet”. And repeating it for seven years may be a sign of honesty, or a sign that the answer has not been found.

The Bias of Weak Markets

Five years of positive return in markets weaker than the United States deserves a more careful reading than it usually gets.

Weak markets have a structural feature: low competition. When competition is low, acquisition costs are low, users stay longer because no strong substitute exists, and per-user margin is higher. Those are ideal conditions for a small company to prove a model.

But the very features that make weak markets profitable are the ones that disappear on entering a strong market. In the United States, esports users have dozens of competing entertainment options: streaming platforms, mobile games, and traditional sportsbooks that already have millions of loyal users and apps pre-installed on phones. Acquisition costs in the United States are higher not only in ad prices but in the number of touches required to convince someone to install one more app.

This is a form of survivorship bias at market level: a model that succeeded in easy conditions does not automatically succeed in hard ones. Doubling a small profitable market does not create a large profitable market. It creates two small profitable markets.

I am not saying this to dismiss the achievement. Five years of positive return is real, and most projects in this industry do not reach it. I am saying it to assign the right weight within the whole picture.

Capital Does Not Build Ecosystems

There is a recurring pattern in global sport that I have tracked for years and that deserves to sit beside this story.

In football, the wave of large money flowing into Gulf leagues over the past few years brought in a string of stars at the end of their careers. Structurally, that money did not build grassroots academies, did not build youth leagues, did not build local supporter culture. It bought image, attention, and a place on the sports tourism map. The ageing star becomes an ambassador, not a seed.

The esports betting industry shows a similar pattern on a much smaller scale. Money flows into tournaments, into teams, into events — but rarely down to the base layer, where young players are trained, where data is properly recorded, where the integrity of small competitions is protected. The result is an ecosystem with a bright surface and a thin foundation.

A betting market that wants liquidity needs a foundation. It needs thousands of small matches with trustworthy results, clean data and stable schedules. If that layer does not exist, everything above it can only serve a small group of users, and every claim about market potential rests on an unverified assumption.

When Commercialisation Runs Ahead of the Foundation

There is another aspect of the same problem, and I see it becoming clearer in how sports platforms build product catalogues.

Women’s competitions in esports and in traditional sport have drawn more sponsor attention in recent years. Looking at budget structures, most of that money attaches to communications campaigns, to corporate social responsibility reporting, to inclusion metrics. It does not attach to stable player salaries, to building development pipelines, to expanding the number of tournaments. When a budget exists mainly to prove something in an annual report, it disappears when the report changes.

The same can happen with new product categories in betting. A platform can add new market types — women’s events, academy events, regional events — as image polish, then cut them when they are no longer needed. For users, those categories appear and vanish without clear reason. For players, it is unstable income.

A healthy market needs product categories that exist because real users are there, not because a page in a report needs filling.

Three Signals to Watch

If I had to pose the right question for this story, it would not be “when will the US esports betting market mature”. That question is too broad to be useful, and it has been asked for seven years without an answer.

A more useful question is: which product structure moves esports viewers from watching to participating, and which structure keeps them there?

From all the evidence above, three signals decide the answer.

The first is quarterly trading volume by market type. If volume grows in simple markets — match winner, total maps — that signals mainstream users entering. If volume grows only in complex markets, that signals a specialist user base, and that base has a very clear ceiling.

The second is regulatory progress in large states. Each state legalising esports betting unlocks a new pool of potential users. But more important than the number of states is how each state classifies the product: if prediction markets are grouped with sportsbooks, ROLR’s cost model changes completely.

The third is user acquisition cost. A company that spends in a measured way will publish that number when it looks good. When it starts rising faster than user lifetime value, the disciplined model comes under pressure, and that is when the real story begins.

Freezing a Moment

There is a line I have kept in my notebook for years of following small matches: vision score never lies, but it also does not know how to tell a story.

Trading volume in a prediction market is the same. It does not lie about how many people took part. It simply does not know how to tell the story of why that number is so small while the stands are full.

Seven years is long enough for a man to become credible, and long enough for an argument to become a habit. What I want to know is not whether Seth Young was right that the market is not there yet. What I want to know is how many times the product changed during those seven years, and when the most recent change was.

Some stars do not choose the spotlight; they simply wait for the right rain. But some stars stand still so long that people forget they were ever able to move.