Google and U.S. Ski Team debut AI-based athlete tool

The new AI-based athlete performance tool from Google and U.S. Ski & Snowboard is set to change how American skiers and snowboarders train. The two organizations announced the collaboration this week. The tool aims to give athletes quicker feedback, stronger safety measures, and improved performance on the mountain. It works in near real time and turns a simple smartphone into a powerful coaching device.

Google Cloud engineers spent months working with the Stifel U.S. Freeski Team and the Hydro Flask U.S. Snowboard Team. They tested the system in harsh winter environments across Austria and Colorado. These sessions helped the engineers understand the long-standing issues coaches face during outdoor training. For years, they had to choose between fast but subjective observation or slow and complex lab analysis. This new approach fixes that problem by combining accuracy with convenience.

The tool uses advanced AI from Google DeepMind to map an athlete’s body in 3D with a single video. Thick jackets, gloves, and helmets usually hide body angles, yet the tool sees through this gear with spatial intelligence. Coaches no longer need athletes to wear sensor suits or markers. A regular smartphone becomes a high-precision motion capture sensor, which makes training more natural and less distracting.

U.S. Ski & Snowboard chief of sport Anouk Patty said this step marks a major moment for coaching. She explained that video always helps coaches, but analyzing it takes too long. The AI-based athlete performance tool changes this by delivering fast breakdowns. It supports safer practice sessions and gives athletes insights that once required specialized labs.

Google Cloud built the system using its full-stack technology. The design includes custom TPUs, Gemini reasoning engines, and research from Google DeepMind. These parts work together to understand human motion with impressive detail. Traditional tools often fail outdoors, but this system stays reliable even in freezing, high-speed settings.

Coaches can record a run from the side of the slope or from the bottom of the course. They upload the clip to a dashboard, where the AI begins its review. The process is quick and smooth, which helps teams respond immediately during training. Athletes also receive conversational feedback. They can ask the tool questions in simple language, such as, “How much faster did the rider need to spin?” The AI then provides clear and useful numbers.

Each session is saved in a central database. Over time, this creates a powerful library of performance data. Coaches can compare past and present footage, track progress, and explore why certain tricks succeed or fail. Because the system supports both freeski and snowboard athletes, it also helps teams study trends from multiple disciplines.

Oliver Parker, vice president of Global Generative AI at Google Cloud, said the project sets a model for global training. He believes the AI-based athlete performance tool proves that high-level coaching can reach more people. If the system can work for elite athletes in extreme weather, he said, it can also help everyday users like physical therapy patients or hobby golfers.

U.S. athletes will continue testing and refining the system as they prepare for future competitions. With the AI-based athlete performance tool, they now gain faster insights, safer routines, and a new edge on the world stage. This partnership signals a future where AI supports human movement in smarter and more accessible ways.