AI transforms U.S. Open experience with real-time analytics for players and fans

The U.S. Open has become a laboratory for artificial intelligence, and the experiment is unfolding in real time across millions of phones. When Coco Gauff played Zeynep Sönmez on a Tuesday night in late August, fans watching through the official app saw something that would have seemed impossible a decade ago: live AI analysis flowing in alongside the match itself, tracking win probability, flagging pivotal moments, and measuring the biomechanical precision of every serve.

The infrastructure behind this is substantial. Cameras positioned around Arthur Ashe Stadium capture the ball, the racket, and the players' bodies in motion—recording data points like elbow flexion, knee bend, and wrist velocity. IBM's Watsonx platform processes this stream of information to generate a "serve quality" score out of 100, a metric that appears in the app after each match. When Gauff won, the analysis noted her "controlled racket preparation and deep knee bend during her setup." By the tournament's end, IBM's technical program director Tyler Sidell said, the system will have generated over a billion data points.

The intent, Sidell explained, is not to replace the unpredictability that makes tennis compelling but to deepen it. "There's so much unpredictability in sports," he said. "We want to provide an insight, but still, let's watch the matches play out." The app reaches roughly 14 million users, and the analytics are designed as conversation starters—ways for fans to engage more deeply with what they're watching without being told what will happen.

But the technology has moved beyond spectating. Players are using it too. Jessica Pegula, who advanced to the fourth round, relies on AI analysis to identify patterns in opponents' serves before stepping on court. "Tennis is a lot of problem solving," she told CBS News. "Serve is a really big one. It's the one controllable shot we have in tennis." Pegula is clear-eyed about the limits: strategy shifts mid-match, players deviate from their tendencies, and instinct sometimes matters more than data. Still, she said, the preparation feels valuable. "It just gives you that feeling of being prepared before you go into a match."

The "likelihood to win" feature has been part of the U.S. Open app for about six years, but this is only the second year fans can watch it fluctuate in real time as a match unfolds. The odds are built from trusted media sources and recent performance records, aggregated to generate a player's chances. When Gauff faced Sönmez, she opened as a 72% favorite—a number that climbed as she took control of the first set. The app's "key moments" feature provided brief analysis as the match progressed: "Gauff has three opportunities to win the match, with Sonmez needing a heroic response," it noted during the second set.

The most dramatic swing came later that evening when Alexander Zverev, the men's No. 1 seed from Germany, played Lorenzo Sonego of Italy in a match that stretched into the early morning hours. Zverev began with an 87% win probability. But as the fourth set wore on, the algorithm shifted sharply: Sonego climbed to 93% likely to win. Then Zverev mounted a comeback, and the odds flipped again. The match became a live demonstration of how quickly certainty can dissolve on a tennis court.

The app also includes Match Chat, an AI assistant that answers questions about players, matches, and the venue itself. Ask it where to find a Honey Deuce—the U.S. Open's signature cocktail—and it returns bar locations and asks which gate you're nearest to. These small conveniences are part of a larger shift: the digital layer has become inseparable from the tournament experience. Fans are no longer just watching tennis. They're watching tennis while consulting a system that watches it too, offering data, context, and prediction in real time. Whether that deepens the experience or simply adds noise remains something each viewer will decide for themselves.