Fifty times a second, cameras at Flushing Meadows are tracking 21 points across a player’s body and racquet. Wrist flex, ball toss, the transfer of energy from the legs up through the torso. By the time the 2026 US Open ends on September 13, the system will have generated roughly 1.2 billion data points, all in service of one question: what makes a great serve great?
The technology is the centrepiece of this year’s collaboration between IBM and the United States Tennis Association, announced on August 24 as the tournament got underway in New York. The two organisations have been building digital products together since 1992, and their work now reaches more than 14 million tennis fans a year through USOpen.org and the US Open app.
It arrives at a moment when sports audiences have made their appetite for this material plain. Second-screen viewing is now the default at major tournaments, and tools that once served broadcasters exclusively, ball tracking, win probability, biomechanical analysis, are being handed directly to anyone with the app installed.
A score for every serve
The headline feature is called Serve Quality, and it runs across all 254 singles matches. Limb-tracking technology developed with IBM’s Bob platform analyses the mechanics of each serve as it happens, measuring efficiency, accuracy, consistency and ball toss. IBM Confluent manages the continuous stream of live data and turns it into a single score generated in near real time.
The point is context. A 130 mph reading on the speed gun tells you about power and nothing else. A Serve Quality score is an attempt to capture technique: whether the toss was where it should be, whether the kinetic chain from legs to racquet actually delivered, whether the motion held up under pressure. It is the kind of judgment a coach makes by eye, produced by software at broadcast speed.
Explaining the match, not just calling it
Two other features fill out the 2026 lineup. Key Moments builds on Likelihood to Win, the probability tool that calculates each player’s chance of victory using live statistics, historical data, expert opinion and match momentum. Where Likelihood to Win says who is ahead, Key Moments explains why, summarising the swings and turning points that flipped a match in an instant.
Match Chat, the tournament’s conversational AI, has been upgraded as well. Fans can ask questions in plain language during a match and get instant answers, some of which now pull in photos and video. The system runs on watsonx Orchestrate, IBM’s agent platform, with AI agents and models trained on the USTA’s editorial style and the particular vocabulary of tennis, so answers read like tennis coverage rather than generic chatbot output.
Fans seem ready for it
IBM published a survey alongside the announcement, conducted by Morning Consult, and the numbers help explain the investment. Among global tennis fans surveyed, 91 percent said they use sports apps during events, and 64 percent expressed trust in AI-powered sports content. Accuracy is the currency here, not novelty. A fan who catches a wrong statistic once will discount everything the app tells them afterwards.
Brian Ryerson, the USTA’s senior director of digital strategy, described the partnership as delivering personalised content and insights to millions of fans while strengthening the organisation’s ability to tell stories across every match. Jonathan Adashek, IBM’s chief global affairs officer, put the emphasis on speed, saying the company can take millions of live data points streaming from the court and route them into workflows that produce usable insights within minutes and hours.
Tennis as a proving ground
Grand Slam tennis has become one of sport’s favourite AI laboratories, and it is easy to see why. The game is measurable in ways football is not: a contained court, two players, discrete points, a clear outcome every few seconds. Wimbledon has trialled AI-generated commentary with IBM, and the Australian Open replaced human line judges with electronic line calling years ago. Systems proven at Flushing Meadows have a way of migrating elsewhere, too. This summer’s World Cup ran semi-automated offside technology built on AI-tracked player skeletons, and camera-based tracking now underpins officiating and analytics across elite sport.
The interesting question is where the data goes next. A serve database with more than a billion entries is a fan feature today, but the same measurements describe biomechanics that coaches, broadcasters and player camps will want for themselves. For now, the scores belong to the fans. For more on how AI is changing sport, visit Mylistingo.
