IBM brings its US Open AI demo back to Madison Square Park

The free September 10th-13th event pairs match screenings with a serve simulator, AI-generated ping-pong recaps and a photobooth beside IBM's Manhattan office.

By · Published

Primary source: IBM Newsroom

Why it matters

IBM is using a free sports activation to make its enterprise AI stack tangible, turning production systems built for the US Open into demonstrations the public can test in seconds.

IBM brings its US Open AI demos back to Madison Square Park — The free September 10th-13th pop-up turns tennis telemetry, match recaps and photo generation into a public showroom beside IBM's Manhattan office.

IBM is bringing a free US Open viewing event back to Madison Square Park, where tennis screenings will share space with public demonstrations of IBM's AI and data technology.

In its August 27th announcement, IBM said the event will run from September 10th through 13, with screenings of the singles semifinals and finals alongside a serve simulator, AI-generated ping-pong recaps and a photobooth that places visitors inside US Open and New York City scenes.

The location puts the event beside IBM's One Madison Avenue office, giving IBM a public setting for demonstrating technology that IBM usually sells through enterprise software, infrastructure and consulting relationships.

AI demonstrations beside the screenings

The three public demonstrations offer a simplified version of IBM's AI pitch. IBM Serve Speed measures a visitor's serve speed and accuracy against professional benchmarks. Calling the Shots uses AI-enabled ping-pong tables to produce a match recap, while the AI Photobooth generates themed images. IBM's announcement does not identify the models running those attractions or explain whether Serve Speed shares technology with the Serve Quality system used during tournament matches.

The park demonstrations are lighter than the production systems IBM and the United States Tennis Association introduced for the 2026 US Open. According to the organizations, those include a personalized Live Updates homepage, AI-generated analysis of important match swings, an expanded conversational Match Chat and the new Serve Quality metric.

Serve Quality tracks 21 positions across a player's body and racquet 50 times per second. IBM and the USTA say the system will process roughly 1.2 billion data points during the tournament, analyzing mechanics such as wrist flex, shoulder rotation, ball toss and energy transfer. IBM says the continuous stream of live data is managed by IBM Confluent, its real-time data streaming platform. Match Chat uses watsonx Orchestrate and, according to IBM and the USTA, a collection of models and agents tuned for tennis terminology and the USTA's editorial style. Those company disclosures give IBM a concrete technical story to tell once visitors move past the photobooth.

IBM applies the "AI-powered" label broadly across the park attractions, while its announcement and the Madison Square Park Conservancy's event listing do not specify what information is collected, how long it is retained or whether visitors can opt out of reuse. Those disclosures would help visitors assess the generated-photo and gameplay attractions before supplying faces, movements or match data for personalized outputs.

Tennis as an enterprise proof point

IBM's work with the USTA predates today's generative AI boom by decades. The organizations have worked together since 1992, and IBM and the USTA say the US Open website and app reach more than 14 million fans globally each year. They supplied that figure without disclosing an independent audience measurement.

IBM presents the longer relationship as a recurring, high-traffic production environment. Tournament traffic arrives in sharp bursts, match data changes after every point and the biggest demand lands during a handful of finals. In IBM's telling, that pressure provides a venue for demonstrating its data, cloud, automation and consulting products together.

An IBM engineering account of the US Open architecture presents the tournament as a case study for IBM's data and cloud systems. It is IBM's own account, but it shows how IBM connects the public-facing fan experience with the infrastructure and software it sells to enterprise customers.

That architecture sits inside the strategy pursued by IBM Chairman and CEO Arvind Krishna. Krishna previously led IBM Cloud and Cognitive Software, served as director of IBM Research and was a principal architect of IBM's Red Hat acquisition, according to IBM's annual-report material. IBM now markets a combination of hybrid cloud infrastructure, AI software and consulting to clients. IBM can present the US Open as a consumer-facing case study with an unusually visible deadline: every point has to appear while fans still care about it.

Four days of tennis and product demos

According to IBM's event schedule, the activation opens from 11 a.m. to 9 p.m. on Thursday, September 10th, with coverage of the women's singles semifinals beginning at 7 p.m. IBM lists Friday's hours as 9 a.m. to 9 p.m., with men's semifinal coverage beginning at 3 p.m.

The same IBM schedule lists Saturday's hours as 11 a.m. to 8 p.m., with women's final coverage beginning at 4 p.m. For Sunday, IBM lists an 11 a.m. to 7 p.m. operating window and men's final coverage beginning at 2 p.m. IBM says admission is free and promises merchandise and giveaways alongside the demonstrations and screenings.

IBM has not disclosed a 2026 attendance target, event budget, conversion goal or data-retention policy for the pop-up. Its format translates the enterprise stack into a serve, a ping-pong recap or a personalized image that takes seconds to understand. The matches provide the crowd, and the park gives IBM a stage for its account of how the underlying AI performs on schedule.

Reader comments

Conversation for this story loads after sign-in.