Kambi Group
Kambi estimates that its AI-powered dynamic live delay technology could deliver 18 million additional accepted live bets annually. The forecast, presented with its third-quarter results on 4 November, puts a number on a familiar frustration: selecting a bet, waiting for confirmation and seeing it fail to go through at the quoted price.1 The commercial opportunity lies in reducing that gap.
The wait has a purpose. Britain’s Gambling Commission explains that sportsbooks delay acceptance to ensure their odds reflect the progress of an event. The appropriate delay depends on the operator’s strategy, the event and the speed of its data.2 During a football match, a goal or red card can change the value of a selection while a customer’s request is being processed.
How Dynamic Acceptance Works
Kambi supplies the technology behind betting operators, including pricing and risk management.3 Its new approach makes the delay responsive to the individual wager. Explaining the change on the results call, chief executive Werner Becher described an assessment of the match situation, the type of bet and confidence in the risk management before deciding whether immediate acceptance is possible.4
Kambi says 59% of live bets traded by its AI system are now accepted with zero-second delay. The broader AI trading system handles 85% of network bets.13 These are different measures: one describes acceptance speed within AI-traded live betting; the other describes how widely the trading system is deployed. The figures show meaningful adoption without establishing that every bet on every partner site receives instant confirmation.
| Measure | Figure | Scope and status |
|---|---|---|
| Additional accepted live bets | 18 million a year | Management estimate; not an achieved total |
| Zero-second live delay | 59% | Within live bets traded by Kambi’s AI system |
| AI trading coverage | 85% | Share of network bets handled by the AI trading system |
| Live NBA bet rejection | 32% lower | Early comparison after the AI switch at the season’s start |
Company-reported figures. The 18 million figure is an annual estimate. The NBA comparison covers an early period; zero-second live delay still involves data transmission and processing.
Zero-second delay also needs a practical reading. Removing an imposed waiting period still leaves data transmission and processing to happen. It cannot make a delayed television picture current. The Commission’s in-play technical standard specifically requires operators to explain that broadcasts are delayed and others may have more recent information.7 A faster confirmation changes one part of the experience, while those information gaps remain.
For the operator, a bet that can be accepted while its price remains valid creates an opportunity that a rejected request might lose. But the 18 million figure is an annual estimate from management, not an achieved total. It counts bets, not unique customers or revenue. The financial outcome will depend on stakes, payouts and costs; some customers whose requests fail may otherwise try again.
There is an early indicator to watch. Kambi’s presentation reports 32% lower live NBA bet rejection following the switch to AI trading at the season’s start on 20 October.1 That is a short observation period. The useful next question is whether the improvement holds across a longer schedule and different match conditions, alongside reliable pricing.
From a Trading Rollout to a Modular Product
Historical partnership · 2024
Tzeract’s Bet Builder partnership with operator kwiff was announced in April 2024.
This capability has a longer history than the latest AI headlines. Kambi’s 2024 presentation traced its AI trading rollout back to the 2022 World Cup. In April 2024, its Tzeract division announced its first commercial partnership with kwiff, providing a Bet Builder product designed to integrate into an existing sportsbook.5 The arrangement showed how an operator with proprietary technology could still buy a specialist AI component.
By July 2026, Kambi reported that its Turnkey Sportsbook network had passed 100 million World Cup bets before the tournament ended. It said both pre-match and live offers were compiled and traded through its proprietary algorithms. That total excluded Odds Feed+ activity.6 The development path has moved through pricing and complex bet combinations into the decisions governing acceptance itself.
The latest financial results provide context for the technology investment. Quarterly revenue increased 23% to €46m, while operating profit rose to €5.1m from €1.6m. Management also cited cost discipline and favourable currency movements.3 Assigning the whole improvement to faster live betting would miss those other influences.
Where Faster Decisions Could Take Betting
Our view is that the next competitive battleground will increasingly be the quality of decisions made while an event is unfolding. Can a system keep a market available, price a complicated selection and accept a bet without introducing avoidable errors? A customer may notice fewer interruptions before noticing any change in the range of markets. That gives operational improvements a route to commercial value.
It also changes the question of what a bookmaker needs to own. The kwiff precedent suggests that control over the customer experience can coexist with buying specialised trading tools. For operators, we see a trade-off between gaining access to a supplier’s models and becoming dependent on its data, availability and development priorities. Owning a platform does not settle every technology decision.
Speed makes the quality of oversight more consequential. A faulty feed or a model that handles an unusual match situation badly could let an error spread through accepted bets quickly. In our assessment, expansion into more sports will make monitoring, testing and the ability to suspend affected markets increasingly important. The relevant measure is reliable acceptance, with errors and outages included in the assessment.
There is a customer-protection dimension too. The Commission expects in-play operators to detect potential harm through information such as changes in betting frequency, time and spending.2 Faster confirmation could increase the pace at which bets are placed. Our view is that investment in acceptance technology should be matched by controls capable of recognising when that pace signals a problem.
Kambi has given the industry a specific claim to test. Future updates can show whether shorter delays produce a sustained improvement in acceptance and commercial returns while pricing and safeguards hold up. Those results will tell us more about AI’s value in betting than the number of products carrying the label.
Prepared with help from ChatGPT.