The tipping point came quietly. For years, in-play betting was the growth story—the exciting new vertical, the feature roadmap priority, the segment operators tracked as a share of handle climbing toward relevance. In 2025, it became the default. Live betting now represents 62.35% of global betting-type market share, up from 59.6% just one year prior, and the trajectory is accelerating at a 13.62% CAGR through 2031. Pre-match is now the minority.
The implications for operators are not subtle. An infrastructure and product stack built around pre-match wagering—with in-play bolted on as a feature—is now structurally misaligned with where the money is. This article lays out exactly what the gap costs, what the leading operators have built, and the sequence in which every operator should prioritize closing it.
Market ShiftThe 50% Threshold Is Gone—Live Betting Is Now the Default
DraftKings reported a landmark in Q1 2025: 54% of its total betting handle was live wagering—the first time in the company’s history that live bets exceeded pre-match bets. This was not a one-quarter anomaly driven by a calendar anomaly or event mix. It reflected a structural shift that Optimove data independently confirmed across the broader US market, where in-play share reached 52% overall.
The US crossing 50% is significant because the American market matured later than Europe. European operators have operated in in-play-dominant environments for years: Greece is at 70% live share, Italy at 57%, Spain at 55%. Even the UK, which tends toward lower in-play volume due to regulatory restrictions on certain bet types, shows a live share of 34% by volume—but critically, 50% by revenue. Across the five major regulated markets of the US, UK, Greece, Italy, and Spain combined, 54% of all bets placed are now live.
| Market | In-Play Share (Volume) | In-Play Share (Revenue) |
|---|---|---|
| Greece | 70% | — |
| Italy | 57% | — |
| Spain | 55% | — |
| United Kingdom | 34% | 50% |
| United States | 52% | — |
| Global average | 62.35% | — |
Sources: OddsMatrix global betting statistics; Optimove live betting insights; Gaming America / DraftKings Q1 2025 earnings.
The global figure—62.35% in 2025 vs. 59.6% in 2024—tells the more important story. This is not saturation; it is acceleration. The market structure around sharp money is also increasingly live-oriented, meaning in-play handle concentration is even higher among high-value bettors than aggregate figures suggest.
Revenue CaseLive Bettors Are Worth Nearly 2x—The Numbers Make the Argument
The share story gets the headlines, but the revenue story is what should drive operator investment decisions. US in-play bettors spend an average of $1,583.90 per month. Pre-match bettors spend $846.20. That is an 87% premium per player—not from a different demographic or a higher-stakes segment, but from the same bettor engaging differently.
The UK data validates this from a different angle. In-play represents only 34% of UK betting volume but 50% of total operator revenue. The yield per live bet is structurally higher—a combination of more frequent event-driven decisions, smaller but more numerous wagers, and greater susceptibility to in-session impulse placement when the product experience is engaging. Every live session is a monetization opportunity that pre-match volume simply cannot replicate.
The aggregate market context sets the stakes. US sports betting hit $165+ billion in total handle and $16 billion in revenue in 2025—a 24% year-over-year increase—with in-play growth cited as a primary driver across analyst coverage. Industry forecasts project US sportsbook revenue exceeding $15 billion by 2027, with in-play expansion as the central growth mechanism.
One Second of Lag Costs You Five Figures Per Event
In-play betting is a real-time market. The moment the odds on a platform lag behind reality by more than one second, arbitrageurs—equipped with faster feeds and automated bet placement—exploit the gap. Industry practitioners estimate this exposure at high five figures per event for operators running delayed odds engines. It is not a theoretical risk; it is a systematic, repeatable cost that compounds across every live event on the calendar.
Leading operators have eliminated this exposure by running proprietary odds engines on edge-computing infrastructure with direct league data feeds, recalculating odds every 200–500 milliseconds. This is not incremental optimization; it is a fundamentally different infrastructure tier. Operators still running batch-updated odds pipelines or relying on third-party feed aggregators with latency buffers are not competing in the same product.
The latency targets differ by sport. Horse racing, where events are short and odds move fast, requires sub-500ms stream latency to keep the video feed and betting markets aligned. Traditional team sports carry a slightly looser threshold—1 to 2 seconds—but any gap beyond that creates the arbitrage window. By 2025, 74% of tier-one sportsbooks have deployed AI-based betting automation, processing over 110 million odds updates per hour. The operators who have not are structurally exposed on every live event.
The watch time case study from a horse racing operator deserves specific attention. Reducing stream latency from 12 seconds to 500ms produced a 29% increase in average watch time per session. Watch time in live betting is not a vanity metric—it is a direct proxy for bet frequency, in-session engagement, and revenue per event. A bettor watching the race is a bettor who can be shown next-race markets, place another bet, and return to the platform. A bettor who closes the stream because the delay makes the experience disjointed is lost revenue.
Micro-Betting Is the Fastest-Growing Vertical—And It Demands New Infrastructure
Micro-betting—wagers on granular in-game events like the next play, next point, or next shot—is not a niche product extension. It tripled in volume during 2023 and has continued compounding. NBA betting volume grew 50% following micro-betting introduction, a figure that indicates the market is not cannibalizing existing handle but expanding the total addressable betting moment within a game.
The product baseline has already been set. 95% of major US sportsbooks now offer same-game parlays—the most widely deployed micro-bet product. Any operator not yet offering SGPs is not competing for a growth segment; they are competing against the industry floor. The ceiling is considerably higher: stream-integrated micro-wagers via Twitch and YouTube Live facilitated over 18 million real-time micro-wagers in the most recent tracked period, a 29% year-over-year increase that signals the convergence of media consumption and wagering intent.
Micro-betting demands infrastructure that most live betting setups were not designed to provide. Traditional in-play requires event-level data resolution at roughly one-to-two-second intervals. Micro-betting requires play-by-play resolution at under 100 milliseconds—a different tier entirely. This requires either proprietary data partnership agreements with leagues (the path taken by DraftKings and FanDuel in the US) or reliance on a managed trading desk with the data infrastructure already in place. Operators without either cannot offer genuine micro-betting; they can only offer delayed approximations that sophisticated bettors will not use.
Build vs. Buy: Choosing Your Live Trading Stack
The infrastructure decision for in-play betting reduces to three tiers: proprietary in-house, managed trading desk (MTD), or hybrid. Each carries different implications for margin, time-to-market, and product differentiation.
Proprietary in-house is what tier-one operators have built. DraftKings, Bet365, and their peers run their own odds engines, data feeds, and risk management infrastructure. The capital requirement is substantial, the build timeline is measured in years, and the competitive moat, once established, is real. This path is not available to most operators on a twelve-month horizon.
Managed trading desks (Sportradar MTS, Genius Sports, Kambi) provide the infrastructure as a service. The scale benchmark is Sportradar’s MTS: 87 billion bet selections processed annually across 1 million matches, with 550+ in-house trading experts managing risk. An operator on an MTD arrangement gets immediate access to this infrastructure in exchange for margin and product control. Time-to-market is weeks rather than years. The trade-off is that product differentiation must come from the front-end experience and personalization layer, not from trading capability.
Hybrid is the path most mid-tier operators should be on: MTD for core sports and event coverage, with proprietary capabilities built incrementally for the highest-volume leagues where the margin trade-off justifies in-house investment.
Mobile is non-negotiable across all three paths. Over 80% of global sports betting activity now occurs on smartphones and tablets across all regulated markets. A live betting product that is not mobile-first is not a live betting product that retains users. The streaming dimension adds another layer: AI-powered broadcast infrastructure has expanded live sports streaming coverage across 240 new channels reaching 390 million additional users. Live streaming is no longer a UX feature—it is a distribution mechanism for the betting window itself.
PersonalizationReal-Time Personalization Is the Next Moat—33% Engagement Lift on the Table
Latency and product depth get operators to parity. Personalization creates differentiation. Operators who connect live odds data to player behavioral signals can identify in-session intent in real time—which sport a player is watching, which markets they have engaged with, how long they have been in session, what their typical bet cadence looks like—and surface relevant suggestions at the moment of highest receptivity.
The engagement case is quantified. Personalized in-play experiences increase user engagement by an average of 33% compared to generic live product experiences. The mechanism is straightforward: a bettor watching a match they care about, shown a market aligned to their betting history, at the right moment in the game, has significantly higher conversion probability than the same bettor shown the same market from a generic odds board.
DraftKings’ product roadmap signals where the market is heading. Per Jefferies analyst coverage, DraftKings is expanding in-play player props and progressive parlays as priority 2025 initiatives—product depth in live markets as the differentiator. An operator with a strong live product but a generic presentation layer will lose on experience even if their odds are competitive. The sharp money research consistently shows that market efficiency differences between operators are smaller than product experience differences in driving player retention.
The technology investment context frames the urgency. The global AI-powered sports betting market was valued at approximately $9 billion in 2024, growing at a 21.1% CAGR toward an estimated $28 billion by 2030. Operator investment is concentrated in three areas: predictive odds-making, personalization and CRM engines, and in-play front-end features. Operators who defer personalization infrastructure investment are not preserving optionality—they are ceding ground to competitors who are building it now.
The practical implementation connects live odds feeds to CRM and player behavior data. An AI betslip recommendation engine ingesting real-time market data alongside a player’s historical bet preferences can surface the right market at the right moment within a live session without requiring manual trading intervention. This is not a conceptual future state; it is the product DraftKings, Flutter, and their peers are deploying now. For operators without proprietary AI infrastructure, the gap is closeable through licensing—but only if addressed in the near term, before the personalization layer becomes another dimension of structural disadvantage.
Action PlanThe Operator Build Checklist: What to Prioritize in 2025–2026
The in-play infrastructure gap is not closed in a single sprint. The following sequence prioritizes by risk mitigation first, then revenue upside, then competitive positioning.
Priority 1: Audit Stream Latency
Benchmark your current odds recalculation frequency and stream latency against the 500ms target for horse racing and 1–2 second target for team sports. Identify the arbitrage exposure window. This is an immediate risk audit, not a product decision—it quantifies what you are currently losing per event before any investment decision is made.
Priority 2: Evaluate Managed Trading Desk
If your in-house odds engine recalculates at intervals greater than 1 second, the path to parity is not building faster infrastructure from scratch. Engaging Sportradar MTS, Genius Sports, or Kambi on a managed trading arrangement closes the latency gap in weeks rather than years, at the cost of margin. For most operators below tier-one scale, this is the correct trade-off.
Priority 3: Launch Same-Game Parlay and Basic Micro-Market Product
95% of major US sportsbooks already offer same-game parlays. If you are in the 5% that do not, this is not a growth initiative—it is table stakes. The baseline micro-betting product must exist before any differentiation strategy is relevant. Launch it, then build toward more granular micro-markets as data partnerships allow.
Priority 4: Connect Live Odds Feed to CRM Personalization Layer
The highest-ROI live betting personalization is not complex to initiate: connect your live odds data to your CRM platform and trigger in-session push notifications based on player sport preference and current market availability. A player watching their preferred team’s match should receive a contextually relevant push at half-time, not a generic promotional offer. This requires an API connection between your odds feed and your CRM, not a rebuild of either system.
Priority 5: Build or License an AI Betslip Recommendation Engine
The personalization ceiling is reached by combining odds intelligence with individual player behavioral signals to surface specific bet suggestions within a live session. This closes the gap between market availability (which every operator has) and player engagement (which most operators undermonetize). Build this capability in-house if you have the data science infrastructure; license it if you do not. The 33% engagement uplift from personalized in-play experiences is too large to leave on the table while building from scratch.
Horizon: Stream-Integrated Wagering Partnerships
The 18 million micro-wagers placed via Twitch and YouTube Live integrations represent the direction of media-betting convergence. This is not a 2025 operational priority for most operators—it requires platform partnerships that take time to negotiate and infrastructure to support. But it should be on every roadmap. The operators who establish these distribution relationships in 2025 and 2026 will have a structural advantage as stream-integrated wagering scales in the years following.
Data Sources & Benchmarks
- OddsMatrix: Sports Betting Statistics 2025 — global in-play share (62.35%), 2024 baseline (59.6%)
- Verified Market Reports: Live Betting Market Forecast — 13.62% CAGR through 2031
- Gaming America: DraftKings Q1 2025 Earnings — 54% live handle share
- Optimove: Live Betting Insights — $1,583.90 vs $846.20 monthly spend, 52% US market share, multi-market breakdown
- LinkedIn / Andrew Mazur: Evolution of In-Play Betting — 1-second arbitrage threshold, five-figure cost per event
- Ververica: Modernizing Sports Betting Technology — 200–500ms odds recalculation frequency
- Dolby Optiview: Low Latency Streaming in Sports Betting — +29% watch time case study, latency targets by sport
- Sportradar MTS: 87 billion bet selections/year, 1 million matches, 550+ trading experts
- Global AI-powered sports betting market: ~$9B (2024), 21.1% CAGR, ~$28B by 2030