Kambi's AI Now Prices 70% of All Bets Across Its Network

Over 70% of all betting activity on Kambi's network now runs through AI. That number comes straight from the sportsbook supplier's trading operation. It's also the clearest sign that machine pricing has moved past the pilot stage and into the core of the product.

Oliver Lamb doesn't dress it up. Kambi's SVP Trading says the models aren't where the advantage lives. The edge sits on more than 15 years of historical trading, pricing and outcome data, and an AI is only as good as the archive under it. That distinction matters. Any supplier can license a model. Very few have a decade and a half of settled markets to train one on.

The timeline tells its own story. Kambi began building AI trading in 2021. By 2022 the technology was proven in live production. By 2024 it sat inside the whole trading operation. This year the system expanded to run several sports at once, and it now prices the majority of bets on the network. Lamb frames that as years of investment. Not a feature bolted onto an existing platform.

What operators get is tighter pricing and better margin protection. Prices are generated and refreshed in real time, with the AI learning continuously from historical and current trading activity alike. Kambi runs a shared network, so its models see trading patterns across many operators and markets at once. No single book can replicate that. Lamb is direct about it: network training models learn from a broad set of trading patterns that one operator, working alone, could never see, detect or build around.

Real-time monitoring sits on top. It flags shifts in model performance, risk exposure and unusual betting behaviour as they happen. For operators the upside goes past holding margin. Fewer rejected bets. Better uptime. Room to launch new products and promotions without gutting profitability.

Product pressure is real. Bet Builders, player props and ever more complex promotional mechanics are far harder to price and risk-manage than a plain single-market book. AI lets Kambi process more signals and refresh prices constantly, so partners can widen customer choice without losing control of exposure. Take dynamic live delays, which set bespoke delays at the individual bet level during in-play. Lamb says this protects margins without slowing the customer experience any more than necessary. At Kambi's network scale, none of it works through manual processes.

Then came the stress test: the FIFA World Cup. Across all 104 matches, Kambi delivered automated trading with 100% uptime and a wide range of combinable options. Lamb points to the difficulty of juggling concurrent global demand, fast in-play swings and player expectations across a B2B network. His summary: automated trading on every match, no downtime at that volume. That, he says, is what AI trading and risk management at Kambi is meant to mean in practice. Not a strategy deck, not a thought experiment.

For affiliates, the read is simple. Pricing engines that update in real time and reject fewer bets convert better, and conversion is the only metric that pays. Partners sending traffic to Kambi-powered books should expect deeper market coverage and more combinable options, which means more click-through opportunities on props and build-your-own slips. The same trading engine and network intelligence used across the World Cup are available to Kambi's partners now, as the foundation for the next major tournament and the product rollouts after it.