US Open's AI scores every serve so its digital team can tell the story behind it
1.2BServe Quality ran across all 254 singles matches of the 2026 US Open; IBM reports 21 tracked points per player at 50 frames a second and roughly 1.2 billion data points over the tournament (CBS News: over a billion, per IBM's technical program director).
“Available across all 254 singles matches, the AI-powered tool uses advanced limb-tracking technology developed with IBM Bob to help analyze the precise mechanics of every serve – with 21 data points across the body and racquet, tracked 50 times per second. That means everything from wrist flex to the transfer of energy from a player’s legs through their torso is captured, generating roughly 1.2 billion data points over the course of the tournament.”
IBM and USTA release, Aug 2026 (also CBS News, Sep 2026), 24 Aug 2026
- Who
- The USTA (United States Tennis Association), organiser of the US Open in New York, with IBM, its digital partner since 1992. IBM says the tournament's app and website reach more than 14 million fans a year.
- Problem
- The serve is the one shot a player fully controls, but fans only see the speed and whether it went in. Explaining the mechanics behind a good serve, for 254 matches over three weeks, is more than a digital editorial team can do by hand.
- Agent
- Cameras around the court track the ball, the racquet and the player's limbs (wrist flex, knee and elbow bend, ball toss). IBM Confluent streams the live data, limb-tracking built with IBM Bob analyses each serve, and IBM watsonx scores efficiency, accuracy and consistency into a Serve Quality number out of 100, with a plain-language summary of what made the serve work. It sits next to Likelihood to Win, Key Moments and Match Chat in the US Open app.
We'll build it for you
Book a callView the prompt
Try it yourself with IBM watsonx
You are the coach or communications lead of a tennis club. Goal: turn the serve statistics from one match into a short, readable summary for a player or for our club's channels. My input: [player first name and level], [match stats: first-serve %, points won on first and second serve, aces, double faults, from our scoresheet or tracking app], [anything the coach noticed, such as ball toss or knee bend], [where it will be published]. Give me: 1) three plain-language observations about the serve, each tied to a number I gave you; 2) one thing to work on next week; 3) a 60-word version for social media. Use only the numbers I provide and say clearly when the data is too thin to conclude anything.What we can do for your club or brand
Most clubs and tournaments already have match data: scoresheets, a tracking app, or a coach's notes. The Court can set up an AI assistant that turns those numbers into short match and serve summaries for your players, parents and social channels, using only your own data, so your coaches and staff spend their time on court instead of writing recaps.
Source:
IBM and USTA release, Aug 2026 (also CBS News, Sep 2026), 24 Aug 2026 · first-party source
What they did
For the 2026 US Open (23 August to 13 September), the USTA and IBM added Serve Quality to the US Open app and USOpen.org. Cameras around the court track 21 points on the player's body and racquet 50 times per second; the data is streamed live and scored by AI into a number out of 100, with a short written explanation of the serve, shown after the match. The same release added a personalised Live Updates home page, Key Moments inside Likelihood to Win, and a Match Chat that can now answer with photos and video.
How it went
IBM reports roughly 1.2 billion data points across all 254 singles matches. The USTA's Senior Director of Digital Strategy, Brian Ryerson, says the AI is "significantly enhancing our ability to tell compelling stories throughout every match": the AI handles the volume of tracking data, and the digital team decides which stories to tell. Limits: figures come from IBM, the tournament's paid partner; CBS News confirms the order of magnitude (over a billion) but no figures on how many fans used Serve Quality have been published; IBM calls the score near real-time, while CBS saw it in the app after the match; the setup needs stadium camera tracking that most clubs don't have.