Kambi: How 15 years of trading data turned into an AI edge

AI has become a buzzword in the sports betting industry over the last couple of years. But for Kambi, AI trading is part of its core sportsbook infrastructure, with more than 70% of betting activity across its network now processed by AI. The technology supports increasingly complex products, from…
AI has become a buzzword in the sports betting industry over the last couple of years. But for Kambi, AI trading is part of its core sportsbook infrastructure, with more than 70% of betting activity across its network now processed by AI. The technology supports increasingly complex products, from Bet Builders to player props and in-play markets, while helping operators improve pricing, risk management and margins. Kambi’s SVP Trading Oliver Lamb tells SBC that the company’s advantage comes from combining AI with more than 15 years of historical trading, pricing and outcome data. After years of development, the technology now underpins its trading operations, including every match of the FIFA World Cup this last summer. AI trading built on years of data Kambi’s AI models draw on accumulated trading data to improve pricing and risk management over time. That’s why Lamb explains that the provider gains an edge that “doesn’t really come from just the AI on its own,” but from building on more than 15 years of historical trading, pricing, and outcome data. Kambi began developing its AI trading capabilities in 2021, with the technology proven in production by 2022. By 2024, AI was fully embedded across its trading operation. This year, the system expanded to handle multiple sports concurrently and now processes more than 70% of bets across the network. Lamb says this reflects years of investment rather than a feature added to an existing product, with the technology becoming core infrastructure for the company. Turning network data into better margins AI allows Kambi to generate and update prices in real time, while continuously learning from historical and current trading activity. For operators, this can mean more accurate pricing, stronger margin protection and greater flexibility to introduce complex products. The network also gives Kambi access to trading patterns across multiple operators and markets, helping it identify anomalies that individual operators might struggle to detect independently. Lamb says: “Network training models here can learn from a broad set of trading patterns that a single operator by themselves couldn’t see, detect, or build around.” Real-time monitoring flags changes in model performance, risk exposure and unusual behaviour as they occur. Alongside protecting margins, this can support fewer rejected bets, stronger uptime and the launch of new products and promotional mechanics without unnecessarily compromising profitability. Supporting more complex betting products Bet Builders, player props and increasingly sophisticated promotional mechanics create greater pricing and risk-management challenges than traditional single-market betting. AI allows Kambi to process more signals and update prices continuously, helping operators expand customer choice without losing control of risk. One example is dynamic live delays, which allow bespoke delays at the individual bet level during in-play betting. Lamb says this can help protect margins without unnecessarily slowing the customer experience. He also explains that these capabilities cannot be delivered through manual processes at Kambi’s network scale. As betting behaviour evolves, the trading engine must respond at the same pace. The World Cup as the ultimate stress test The FIFA World Cup tested Kambi’s AI trading under sustained, high-volume conditions. Across all 104 matches, the company delivered automated trading with 100% uptime, alongside a broad range of combinable betting options. Lamb highlights the challenge of managing concurrent global demand, rapid in-play changes and player expectations across a B2B network: “Automated trading across every match, no downtime at that volume. That’s what AI trading and risk excellence at Kambi is meant to mean in practice, not just a strategy deck or a thought experiment,” he says. For Kambi, the tournament demonstrated the capabilities of an infrastructure developed over several years. The same underlying trading engine and network intelligence are available to its partners, providing a foundation for future major sporting events and continued improvements in product choice, risk management and commercial performance.