Skip to content
Operators

There Are Hundreds of Sportsbooks. Who Actually Makes the Odds?

Official data providers now supply pricing, risk and bet-acceptance services once handled in-house, reshaping who really controls a sportsbook's odds.

Industry Analyst & Commercial Partnerships

· 13 min read

There Are Hundreds of Sportsbooks. Who Actually Makes the Odds? — by Martin Eriksen, Industry Analyst & Commercial Partnerships · Gambler Media

Behind the sportsbook logo sits a largely invisible chain of official data rights, pricing models, trading systems and risk engines. We followed a bet backwards to see how much of modern bookmaking is actually done by the bookmaker.

Open five sportsbooks before a Premier League match, and they look like five competing bookmakers. The logos are different, the promotions are different, and the odds will probably differ slightly too.

Then the match starts. A player is sent off and, within seconds, the match odds move, the next-goal market changes, player props are repriced and dozens of bet-builder combinations suddenly look different.

It is natural to imagine traders sitting inside each bookmaker, reacting independently to what just happened. Sometimes that is part of the story. Increasingly, it is nowhere near the whole story.

The information might have come from an official data feed controlled by one company. A separate model may have calculated the new probabilities. Another system may be managing liabilities and deciding how much exposure the sportsbook is prepared to take. Some of that might belong to the operator whose logo you see. Some of it might not.

The modern sportsbook is becoming less like a factory that makes every part of its product and more like a company deciding which parts of bookmaking are important enough to own.

That makes a seemingly simple question surprisingly difficult:

Who actually makes the odds?

One Bet, Several Companies

Start with the event itself.

Before anyone can decide whether Liverpool should be 1.80 or 1.85, the betting system needs to know what is happening on the pitch. For pre-match betting, a few seconds usually make little difference. In live betting, they can matter enormously.

In English football, Football DataCo holds the official data rights for competitions including the Premier League and the EFL. Genius Sports has the exclusive right to collect and distribute that official low-latency data to sportsbooks through the 2028/29 season.

According to Genius Sports’ agreement with Football DataCo, it captures data from more than 4,000 matches a season and distributes it to hundreds of licensed sportsbooks in under a second. From the 2025/26 season, the exclusive agreement also includes player-market data used for bets involving shots on target, assists, passes and other statistics.

4,000+

Matches per season from which Genius Sports captures official data

550+

Sportbook brands with relationships with Genius Sports at year-end 2025

800 / 900

Sportsbook clients and media companies in Sportradar’s MLB data distribution network

Genius Sports and Sportradar company filings

That means the betting supply chain can begin before an odd exists. Someone owns the commercial rights to the information, someone collects it at the stadium, and someone distributes it quickly enough for the betting market to react.

The same structure can be seen in American sports. Genius is the NFL’s exclusive distributor of official real-time play-by-play, Next Gen Stats and official betting data. DraftKings, despite having extensive in-house modelling and trading capabilities, still buys official data and related products from Genius through agreements such as its NFL official-data partnership.

Sportradar has a similar arrangement with Major League Baseball. Under its agreement running through 2032, it exclusively distributes MLB’s ultra-low-latency official data and Statcast content across a network that includes around 800 sportsbook clients and 900 media companies. MLB also received an equity interest as part of the expanded partnership.

The leagues are therefore much more than the events people bet on. Their data has become a valuable asset within the betting industry itself.

The Business Before the Odds

The amount being spent on that invisible layer gives some sense of its importance.

Genius Sports reported $246.5 million in data and streaming rights costs in 2025, up from $188.1 million a year earlier. Its filing says the increase was driven primarily by its strategy of securing official data rights.

The same Genius Sports 2025 annual filing says the company had relationships with more than 550 sportsbook brands at year-end, with services ranging from live data feeds through in-game oddsmaking and risk management.

Sportradar spent even more. Its 2025 accounts recorded €404.3 million in sports-rights expenses, around €52 million more than the previous year. The company attributed much of the increase to deals including ATP tennis, its renewed MLB partnership and additional rights acquired through IMG ARENA.

These are extraordinary sums for something the average bettor hardly notices.

We spend plenty of time comparing whether one bookmaker is offering 2.05 and another 2.10. Much less attention is paid to the commercial battle taking place before those prices reach the screen.

And speed is part of what is being bought.

At its 2025 investor day, Sportradar described data travelling from venue to bookmakers at roughly 500 milliseconds. For live betting, that is not simply a technical boast. A stale price can quickly become a bad price when something significant happens in the match.

~500 milliseconds

Time for data to travel from venue to bookmakers, per Sportradar’s 2025 investor day

65 million

Active bettors whose activity is managed through Sportradar’s Managed Trading Services

Sportradar

So one of the first competitive advantages in modern bookmaking may have nothing to do with a bookmaker’s app.

It can access the right information quickly enough.

Who Turns the Event Into a Price?

Data still isn’t an odd.

Someone has to take what is happening and estimate what it means for the probability of every affected outcome.

That job used to be strongly associated with the bookmaker’s trading room: odds compilers, traders and risk managers building and moving markets.

Those people still exist. But a surprisingly large amount of the work can now be bought from outside.

Genius describes its sportsbook offering as including automated oddsmaking, specialist traders and technology capable of creating events, setting odds, changing them while a match is being played and settling markets afterwards. Sportsbooks can take as much or as little of this as they need.

Sportradar goes further down the chain through its Managed Trading Services business. Its 2025 filing describes AI and machine learning being used to adjust odds according to liabilities, manage sportsbook risk and support bet-acceptance decisions for individual operators. It also offers different trading strategies so two operators using its technology do not have to behave identically.

The accounting description of that service is particularly revealing. Sportradar’s clients can forward proposed bets, or bet slips, to the company for an assessment of whether the wager should be accepted. Sportradar can accept or decline the bet slip and, under some agreements, shares in the resulting revenue or loss. This is described directly in Sportradar’s SEC filing.

That is a long way from merely supplying a football score.

An external supplier can potentially be involved in: data → odds → liabilities → risk → bet acceptance.

And the scale is much larger than most customers would probably guess.

Sportradar says its Managed Trading Services operation handled around €35 billion of betting turnover in 2024, accepting 9 billion bets and managing activity from around 65 million active bettors.

Nobody logs into a Sportradar sportsbook to place those bets. They are betting with the operators that use Sportradar underneath.

That is precisely why the number is interesting. There is an enormous bookmaker-like operation sitting behind consumer sportsbook brands that many bettors have never heard of.

You Can Buy Most of a Bookmaker

Once you start looking at the industry this way, the idea of launching a sportsbook changes considerably.

A company does not necessarily have to build an international trading operation capable of pricing football, basketball, baseball, tennis, table tennis and hundreds of smaller competitions around the clock.

It can buy one.

Kambi’s core Turnkey Sportsbook is essentially built around that proposition. Operators can purchase a sportsbook platform, odds, trading and risk-management capabilities rather than developing every component themselves.

Sportradar likewise offers a betting and gaming platform alongside modular data, odds and managed trading. Genius provides customised sportsbook solutions ranging from live-data feeds and in-game oddsmaking to risk management.

There is another layer that makes the supply chain even less obvious.

Genius’s customer base has included not only consumer-facing operators such as bet365, DraftKings, Flutter, Entain and Fanatics, but major B2B gaming technology providers including Kambi and OpenBet, according to its SEC filings.

In other words, the company supplying technology to your sportsbook may itself be purchasing a component from another supplier.

A simplified Premier League bet could therefore travel through something like:

Football DataCo → Genius Sports → sportsbook/trading technology → bookmaker → bettor.

That does not mean every bet follows that path, and arrangements differ substantially between operators and sports.

But it explains why the question “who makes the odds?” quickly stops having a single-company answer.

The Same Supplier Does Not Mean the Same Odds

There is an obvious conclusion to draw from all this, and it is the wrong one.

If competing sportsbooks use the same underlying suppliers, perhaps their odds are effectively the same product with different logos.

Not necessarily.

A base probability or externally supplied price is only one input into what eventually appears on the screen.

Different operators can choose different margins. They can have different liabilities because their customers have bet differently. Their tolerance for risk can vary. One bookmaker may be happy to accept a large position while another wants to move the price. Trading strategies can differ by market, country, customer segment or sport.

Sportradar specifically markets tools that optimise odds for individual bookmakers based on their own turnover flow and liabilities, while its managed-trading product supports bespoke pricing strategies and odds differentiation between operators.

Kambi has been moving in the same direction with modular products that give operators more control over which externally traded prices they use.

So two sportsbooks might begin with the same official match data, or even the same external odds feed, and still end up showing different prices.

Shared infrastructure is not the same thing as a shared sportsbook.

This is probably the most important qualification in the article. The interesting story is not that hundreds of brands are secretly identical. It is that the underlying industrial structure is much more interconnected than the customer-facing brands suggest.

Why Build It Yourself, Then?

If specialist suppliers can price enormous numbers of markets at scale, the obvious question is why a large operator would ever spend the money to bring those functions in-house.

DraftKings provides a particularly clean answer.

In 2020, its own filings described a sportsbook made from both proprietary and third-party technology. DraftKings controlleaspects such asng its user experience, promotions, account manageme,nt and compliance, while an integrated Kambi platform provided the betting markets, od,ds and risk management.

The DraftKings brand was prominent. A substantial part of the traditional bookmaking underneath it was being provided by somebody else.

Then DraftKings bought SBTech.

Its filings at the time explicitly described bringing more of the sportsbook stack under its control, migrating away from Kambi and becoming a more vertically integrated operator.

By 2026, the description is very different. DraftKings now says it originates prices and manages risk in its sportsbook, with hundreds of data scientists and machine-learning engineers building sports models and a dedicated trading desk fine-tuning live prices.

Yet DraftKings still purchases official data and other specialist products from outside companies.

That isn’t a contradiction. It is probably the best illustration of where the industry is heading.

The decision is not simply:

build or buy.

It is:

What is important enough for us to build, and what are specialists still better placed to supply?

For a large operator, proprietary pricing on major NFL or NBA markets might provide meaningful differentiation. Owning the trading technology may allow faster product development, more control over margins and fewer dependencies on another company’s roadmap.

But recreating every data feed, model and niche market internally would be a very different proposition.

Even “In-House” Doesn’t Really Mean In-House

LeoVegas makes the point particularly clearly.

The company is gradually migrating from Kambi’s Turnkey Sportsbook to its own proprietary sportsbook platform.

If the story stopped there, it would fit the familiar narrative: successful operator becomes large enough and replaces its supplier with internal technology.

Except LeoVegas simultaneously signed a longer-term deal for Kambi Odds Feed+.

Its own sportsbook can now select from Kambi’s library of externally traded odds through an API. The agreement explicitly says the feed is designed to complement LeoVegas’s in-house offering while its migration away from the full turnkey product continues. Kambi and LeoVegas Odds Feed+ agreement

That is almost the perfect illustration of how blurry “in-house” has become.

LeoVegas can own its sportsbook without insisting that its employees personally price every market it offers.

The logic is easy to understand. Perhaps an operator believes football in its most important markets is strategically important enough to model and trade internally. But does it also need to build a world-class pricing operation for lower-volume basketball competitions, table tennis, Korean baseball and dozens of other sports?

A modular supplier allows an operator to own the parts where it believes it can differentiate and purchase the long tail elsewhere.

That is much closer to modern software businesses generally than to the traditional image of a bookmaker.

The Trader You Don’t See

There is another reason the old picture of the bookmaker is becoming less accurate.

A growing proportion of the trading is no longer being performed manually at all.

Kambi reported that 48% of bets across its network in 2025 were priced and traded by AI. By the first quarter of 2026, it said the figure had risen to more than 60%.

The qualifier matters: that is Kambi’s network, not 60% of all sports betting.

But it is still a remarkable number.

A bettor looking at a price does not know whether it was created by a human trader, an operator’s proprietary model, an external supplier’s algorithm, an AI trading system or several of those working together.

In our previous look at how AI might change the bettor’s side of the experience, the question was whether software will increasingly sit between the customer and the sportsbook.

Behind the sportsbook, much of that transition has already happened.

AI is not waiting for somebody to launch an “AI bookmaker”. It is quietly becoming part of the machinery that produces the existing one.

Scale Is Part of the Product

There is a temptation to regard outsourced bookmaking as the cheaper or less sophisticated alternative to having your own traders.

That is too simple as well.

A specialist supplier sees activity across a much larger network than most individual bookmakers ever could.

Sportradar says its MTS systems processed 87 billion bet selections in 2024. It argues that this scale of real-time betting and liquidity data improves its ability to price risk for individual operators.

The supplier may therefore know less about a particular bookmaker’s brand and customers than the bookmaker itself, but considerably more about how an enormous range of markets behave globally.

This creates a real trade-off.

Owning the technology gives an operator control and the possibility of genuine differentiation. Using a large network gives it access to specialist expertise and data generated from billions of bets.

Neither approach is automatically superior.

The largest operators increasingly appear to combine the two.

So What Does the Bookmaker Actually Own?

After following the chain backwards, the word bookmaker begins to look surprisingly broad.

One sportsbook may own its pricing models, trading and risk technology but buy official data. Another may own the customer-facing platform while outsourcing most of the trading. A third may use a turnkey sportsbook almost end to end.

Another can migrate to a proprietary platform while continuing to purchase an external odds feed for selected markets.

The UK Gambling Commission effectively recognises this complexity. Its guidance notes that gambling software can involve “complex and extended supply chains”, with different businesses carrying out one or more parts of manufacture, supply, installation and adaptation. It also notes that multiple parties can each develop components which, together, create the final gambling product. UK Gambling Commission: What is gambling software?

At the same time, operators cannot simply outsource their regulatory responsibility along with the technology. The Commission makes clear that licensees remain responsible for ensuring third-party arrangements comply with their obligations. UK GC: Licensees’ responsibilities for third parties

So outsourcing technology does not mean outsourcing responsibility.

It also doesn’t make the consumer-facing operator irrelevant.

The bookmaker still decides which products to offer, how to present them, how much margin to take, what limits to apply, how to acquire customers, how to handle promotions and, depending on its model, how much of the underlying pricing and risk it wants to control.

Those choices may increasingly be where the distinction between bookmakers lies.

That leads to a different way of looking at the hundreds of sportsbook brands competing for the same customers.

From the outside, they appear to be separate destinations. Underneath, they can be participants in the same much larger ecosystem of leagues, official data holders, technology providers, models and trading systems.

The interesting question, therefore, isn’t really whether Kambi, Sportradar, Genius Sports or the bookmaker “makes the odds”.

Often, several of them have contributed something before the customer ever sees the price.

The more revealing question is:

Which parts of being a bookmaker does your bookmaker believe are important enough to own?

For some, the answer is almost everything. For others, it may be surprisingly little.

And somewhere between those two extremes is where much of the modern sportsbook industry now operates.

Contribute