UFC and IBM Build AI Tools for Fight Analysis
The partnership announced by the Ultimate Fighting Championship and IBM in November 2024 has developed into an AI-powered system for turning historical fight data into broadcast and digital insights. The original announcement projected an early 2025 debut; later IBM reporting documents a staged rollout, with pre-fight analysis first and live in-fight insights added in November 2025.
How the Insights Engine works
The system draws on the promotion’s historical fight data. IBM says an editorial dashboard lets staff query statistics and identify storylines; IBM’s Granite models help draft natural-language copy for broadcast and digital use. Editors check the relevant statistics and shape the final narrative, so the tool supports the production team rather than independently judging fights.
IBM has reported that the system helped the content team produce three times as many insights while reducing the time needed to generate them by about 40%. Those are company-reported results for the editorial workflow, not a measure of fight accuracy or audience growth.
From pre-fight stories to live moments
The first phase generated pre-fight context and storylines from the sport’s data. In November 2025, the partnership introduced In-Fight Insights, designed to flag milestones, streaks and records during live events. IBM reported that the first on-screen example came at the November 15, 2025 event at Madison Square Garden, when Valentina Shevchenko reached 60 career takedowns—then a record for a woman in the promotion.
The official descriptions focus on verified statistics and contextual storytelling. They do not describe a system that predicts fight winners. That distinction matters: the value of the partnership is helping staff find and explain relevant data faster while the action unfolds.
Sources
IBM: Partnership announcement and early 2025 plan
IBM: How the Insights Engine works and its first live display
IBM: In-Fight Insights launch announcement
IBM: Partnership and platform overview