Introduction
A sportsbook looks simple from the outside: take bets, pay winners, keep the margin. In practice it is one of the hardest gambling businesses to run well. The product depends on prices that move every second. The margin swings with results you cannot control. And in most regulated markets the tax bill is rising faster than the revenue.
It can also be a very good business. US sportsbooks kept a record $16.96bn of their customers’ bets in 2025. Well-run operators with a loyal player base are valuable assets, and buyers compete for them.
This guide explains how to start one in 2026: the ways to launch, what the platform needs, what it costs, how the economics work and what can go wrong. It is written for operators and investors, not bettors.
The short version:
- Most new sportsbooks launch on a white-label or turnkey platform. Building your own costs $2m or more and takes 12 to 24 months.
- Margin, called hold, is the number that matters most. US hold rose to about 10.2% in 2025, largely because of parlays.
- Betting taxes are rising across the US and in the UK, Brazil and elsewhere. Model the tax that will apply, not the one that applies today.
- In the US, prediction markets such as Kalshi have become serious competitors to licensed sportsbooks.
- Buying an existing sportsbook with a trading history is often faster and less risky than launching from zero.
The Sports Betting Market in 2026
The US is the best-documented market, and its 2025 numbers show where the business is heading.
| US commercial sports betting | 2025 | Change |
|---|---|---|
| Handle (total amount wagered) | $166.94bn | +11% |
| Revenue (amount kept after paying winners) | $16.96bn | +22.8% |
| Implied hold | About 10.2% | Up from 9.2% in 2024 and 6.9% in 2019 |
| State taxes paid | $3.71bn | +32.4% |
Two things stand out. Revenue grew twice as fast as betting volume, because sportsbooks are keeping a bigger share of each dollar. Parlays and same-game bets carry much higher margins than single bets, and they now make up a large part of US handle. And tax grew faster than revenue, because states keep raising rates.
Growth in betting volume has slowed. US handle fell about 2% in the first quarter of 2026, the first quarterly decline on record. Part of the reason is a new kind of competitor.
Prediction markets
Platforms such as Kalshi and Polymarket offer sports “event contracts” regulated federally by the CFTC rather than by state gaming regulators. To the customer they look like bets. Kalshi’s sports business was estimated at about $1.3bn of annualised revenue in early 2026. The AGA estimates that more than $500m of potential state sports betting tax has moved to these platforms. DraftKings, FanDuel and Fanatics have all launched prediction products of their own, and several states are fighting the platforms in court.
Outside the US, Brazil became one of the world’s largest betting markets in its first year of regulation in 2025. In Europe, the UK, Italy and Spain remain the big regulated markets, with the UK’s online betting duty rising from 15% to 25% in April 2027. Our guide to starting a sports betting business in Brazil covers that market in detail.
Three Ways to Launch
There are three ways to get a sportsbook live, and a fourth that many people overlook.
White label. A provider runs the platform, the odds feed, the trading and often the payments. You supply the brand and the players. It is the fastest and cheapest way to start, but the provider takes a large share of revenue and you control very little of the product.
Turnkey. You license a complete platform but run more of it yourself, usually under your own licence. Setup costs more and the revenue share is lower. You get more control over the product, pricing and data.
Build your own. Only for well-funded teams with sportsbook experience. You own everything, but you need developers, traders and data contracts before you take a single bet.
| Model | Typical setup cost | Ongoing cost | Time to launch | Best for |
|---|---|---|---|---|
| White label | About $5,000–$30,000 | Roughly 10–30% of GGR, sometimes plus a monthly fee | 1–3 months | First brand, affiliates with their own traffic |
| Turnkey | About $15,000–$50,000 for the platform; $100,000–$500,000 with operations included | A lower revenue share or fixed fees, plus odds feed and payments | 2–4 months | Operators who want control and have a licence |
| Own build | $2m+ | Your own team, plus data licensing that can run $300,000–$1.2m+ a year for full coverage | 12–24 months | Established groups with scale |
These are market ranges from industry sources, not quotes. The cheapest headline price often hides the most expensive contract. Ask whether the revenue share is taken on GGR or on revenue after bonuses. Ask whether the odds feed is included for in-play on every sport you want. And ask whether you can take your player data with you if you leave.
The fourth option is to buy an existing sportsbook. A brand with a licence, a trading history and an active player base removes most of the launch risk. You pay for proven numbers instead of building them. Current sportsbooks and casinos for sale are listed on our marketplace.
What a Sportsbook Needs to Run
Whichever model you choose, the same building blocks have to work together.
Odds, trading and risk management
This is the core of the business. The odds feed supplies prices for pre-match and live markets. The trading team or service adjusts those prices, sets limits and decides how much liability to carry on each event. Risk tools flag sharp bettors, arbitrage, bonus abuse and suspicious betting patterns.
Many smaller operators use a provider’s managed trading service rather than hiring traders. That is a sensible starting point. But it means someone else is setting the prices that decide your margin, so read the service levels carefully.
Player account, wallet and payments
The player account platform handles registration, KYC checks, the wallet, limits and responsible-gambling tools. Payments need local methods for each market you serve. Withdrawals have to be fast: slow payouts lose sports bettors quicker than almost anything else.
Product
The product has to match the market. Football dominates in Europe, Latin America and Africa. American sports dominate the US. Cricket matters in India and parts of Asia, and tennis and basketball everywhere.
Beyond sports coverage, players now expect:
- Live betting with fast, stable prices and quick settlement.
- Bet builders and same-game parlays, which combine several markets from one match. They are also the highest-margin bets a sportsbook offers.
- Cash-out, so players can settle a bet early.
- Esports and virtual sports to fill quiet periods in the sporting calendar.
Marketing, CRM and bonuses
Sports betting is a promotion-heavy business. You need a bonus engine that can run free bets, odds boosts and loyalty offers, and a CRM that can target them. You also need an affiliate system. Watch the cost: in high-tax markets operators have been cutting promotions to protect margin.
People
Even on a white label you need people for customer support, payments, compliance and AML, marketing, and someone who understands trading well enough to challenge the provider.
Licensing and Betting Tax
The licence you need depends on where your players are, not where your company is. Offshore licences such as Curaçao or Anjouan cover markets without local licensing. The US states, the UK, Brazil, most of Europe and Ontario each require their own licence. Our guide to the top gambling licences compares the options, costs and timelines.
Tax is the other half of the decision. Sports betting is usually taxed on gross gaming revenue, and rates have been going one way.
| Market | Online sports betting tax | Recent change |
|---|---|---|
| France | 59.3% of GGR in public levies | Up from 54.9% in July 2025 |
| New York | 51% of GGR | Unchanged; the highest in the US |
| Illinois | Graduated, up to 40% of GGR, plus 25–50 cents per bet | Per-bet fee added July 2025 |
| Netherlands | 37.8% of GGR | Up from 34.2% in January 2026 |
| Pennsylvania | 36% of GGR | Unchanged |
| United Kingdom | 15% of profits | Rising to 25% for remote bets in April 2027 (UK horse racing stays 15%) |
| Louisiana | 21.5% of GGR | Up from 15% in 2025 |
| Ohio | 20% of GGR | Doubled from 10% |
| Maryland | 20% of GGR (mobile) | Up from 15% in 2025 |
| New Jersey | 19.75% of GGR (online) | Up from 13% for FY2026 |
| Brazil | 13% of GGR in 2026 | Rising to 14% in 2027 and 15% in 2028 |
High taxes change how a sportsbook behaves. In Illinois, the per-bet fee led several operators to add service charges, and the number of bets fell. In Pennsylvania and Ohio, operators cut promotions. When you build the business plan, use the tax rate that will apply in year two and three, not just the current one.
The Economics: A Worked Example
Two numbers decide whether a sportsbook makes money: hold and tax. This illustration shows how sensitive the business is to both.
Assumptions, for illustration only: a sportsbook taking EUR 20m of bets a year on a white-label platform. Bonuses and free bets cost 15% of GGR, and the platform takes 15% of GGR. Payments and KYC cost EUR 80,000. Marketing and affiliates cost EUR 500,000, and staff and overheads EUR 300,000.
| EUR | 8% hold, 15% tax | 10% hold, 15% tax | 10% hold, 25% tax |
|---|---|---|---|
| Handle | 20,000,000 | 20,000,000 | 20,000,000 |
| GGR (handle × hold) | 1,600,000 | 2,000,000 | 2,000,000 |
| Bonuses and free bets | 240,000 | 300,000 | 300,000 |
| Betting tax on GGR | 240,000 | 300,000 | 500,000 |
| Platform revenue share | 240,000 | 300,000 | 300,000 |
| Payments and KYC | 80,000 | 80,000 | 80,000 |
| Marketing and affiliates | 500,000 | 500,000 | 500,000 |
| Staff and overheads | 300,000 | 300,000 | 300,000 |
| EBITDA | 0 | 220,000 | 20,000 |
Three lessons come out of it.
First, two points of hold is the difference between breaking even and making money on the same betting volume. Hold depends on your product mix as well as luck. More bet builders and parlays mean higher hold, while more single bets on popular football matches mean lower.
Second, a tax rise can wipe out the profit. Moving from 15% to 25%, which is what UK remote betting faces in April 2027, takes this business from EUR 220,000 to almost nothing.
Third, fixed costs need volume. Marketing and staff do not shrink when hold is low. Small sportsbooks need either low-cost traffic, such as an affiliate’s own audience, or a niche where they can price better than the big brands.
Risks Specific to Sportsbooks
A casino’s margin is set by the maths of its games. A sportsbook’s is not, and that creates risks casino operators rarely face.
- Results risk. A run of favourites winning can turn a good month into a loss. Small books with concentrated liability on big matches feel this most.
- Sharp and professional bettors. Players who beat your prices, arbitrage between books or exploit slow live odds can drain margin. Good risk tools and sensible limits matter from day one.
- Bonus abuse. Free-bet and odds-boost offers attract organised abuse. Weak controls make promotions far more expensive than budgeted.
- Integrity. Match-fixing and suspicious betting, especially in lower leagues and esports, carry both financial and regulatory risk. Licensed operators are expected to report it.
- Seasonality. Revenue follows the sporting calendar. Summers without major tournaments are quiet in most football markets.
- Data and official feeds. Some leagues and regulators push operators towards official, paid data, which raises costs.
- New competition. In the US, prediction markets now offer sports contracts with lower margins than traditional sportsbooks, and they are promoted almost like betting brands.
None of these are reasons not to start. They are reasons to budget for volatility, keep reserves, and choose a trading set-up you understand.
How Sportsbooks Are Valued When You Sell
If you build a sportsbook well, it becomes a sellable asset. Buyers look at it differently from a casino, mainly because the margin moves so much from year to year.
- Normalised hold, not last year’s hold. A buyer will average hold over two or three years, or compare it with market norms for your product mix. A lucky season does not raise the price, and an unlucky one should not cut it too far.
- Handle and active players. Steady betting volume and a stable base of active players count for more than one strong quarter.
- Margin quality. Revenue built on bet builders and parlays is valued differently from revenue built on heavily promoted single bets.
- Tax going forward. The buyer rebuilds the P&L at the tax rates that will apply after closing, including any legislated increases.
- Platform and licence. A white-label contract that can be ended on a change of control, or a licence held by the provider rather than by you, will reduce the price.
The same discipline that makes a sportsbook profitable also makes it easier to sell: clean data, a licence in your own name, and a player base that is not built entirely on bonuses. A valuation is a good way to see where you stand before you go to market, and our due diligence guide covers what buyers will ask for.
Frequently Asked Questions
How much does it cost to start a sportsbook in 2026?
A white-label launch can start from around $5,000–$30,000 in setup fees, plus a revenue share of roughly 10–30% of GGR. A turnkey platform with operations typically runs $100,000–$500,000. A custom build costs $2m or more. On top of the platform you need a licence, working capital for marketing, and reserves to cover losing months.
What is the difference between white-label and turnkey sportsbooks?
With a white label, the provider runs most of the operation, often including trading and payments, and takes a larger revenue share. With turnkey, you license the platform and run more of it yourself, usually under your own licence, for a lower share and more control.
What is a good hold for a sportsbook?
It depends on the product mix. US sportsbooks held about 10.2% of handle in 2025, helped by parlays, which hold far more than single bets. Books that mostly take single bets on major football leagues typically hold less.
Do I need a licence to run a sportsbook?
Yes. You need a licence that covers the markets your players are in. Many markets, including the US states, the UK, Brazil and most of Europe, require their own local licence.
How long does it take to launch a sportsbook?
One to three months on a white label once the licence and payments are in place. Two to four months on turnkey. Twelve to 24 months for a custom build. Licensing can add several months in regulated markets.
Are prediction markets a threat to sportsbooks?
In the US, yes. CFTC-regulated platforms such as Kalshi and Polymarket offer sports contracts across most states, including those without legal sports betting, at lower margins. Major sportsbooks have responded by launching prediction products of their own. Outside the US, the effect so far is limited.
Is it better to start a sportsbook or buy one?
If you have traffic and patience, starting can work. If you want a licensed business with proven hold and active players from day one, buying an existing sportsbook is usually faster and less risky.
Conclusion
Starting a sportsbook in 2026 is less about software than it used to be. Good platforms are widely available. The hard parts are the ones the software does not solve: pricing well enough to protect your hold, surviving rising taxes, acquiring players at a sensible cost and holding on to them.
If you are launching, start with the market and its tax rate, then pick the platform model that fits your capital and team. Build the business plan around a realistic hold and the tax rate for year three, not today’s. If you would rather skip the launch risk, look at buying a sportsbook that already has a licence, a trading history and players.
Thinking about buying or selling a sportsbook? Email gabriel@casinosbroker.com for a confidential conversation.
This guide is general commercial information, not legal, tax or investment advice. Gambling should only be offered to adults, in line with local law.
Get new deals before they go public
New casino, sportsbook and affiliate listings go to subscribers 48h before we list them publicly. One email a week at most.
Prefer Telegram? Join iGaming professionals on @igamingdealflow
