U.S. Ski & Snowboard, the national governing body for skiing and snowboarding in the United States, and Google have announced a collaboration on an industry-first AI video-analysis tool built on Google Cloud for skiing and snowboarding. Led by the high-performance needs of world-class athletes, including U.S. Olympians, this experimental AI tool is designed to enhance training precision and mountain safety by delivering near real-time, data-driven insights.
To develop the AI tool, Google Cloud engineers worked side-by-side with the Stifel U.S. Freeski Team and Hydro Flask U.S. Snowboard Team in extreme environments in Austria and Colorado to tackle a long-standing data blind spot in winter sports. For decades, elite coaching has been a choice between two extremes: subjective human observation on the mountain or high-precision data inside a laboratory. Traditional motion-capture requires athletes to wear specialised suits covered in sensors, which are impractical for outdoor training and often fail in sub-zero, high-velocity conditions. This AI tool replaces fragile, wearable hardware with cutting-edge AI from DeepMind that can map a human body in 3D using only video. By leveraging spatial intelligence to “see” through bulky winter gear without the need for sensors.
“This new AI tool is a major coaching development,” said Anouk Patty, Chief of Sport, U.S. Ski & Snowboard. “Video is the most commonly used and effective coaching tool, but analysing it used to be a manual, time-consuming process. By working to create an AI tool that works with a smartphone, we can now analyse competition-grade footage with an extra layer of insight. This isn’t just about winning medals; it’s about providing our athletes with the safest and most advanced training environment in the world.”
The tool is built with Google Cloud’s full-stack AI capabilities to achieve near real-time analysis on the mountain. From custom TPUs in Google data centres, to the reasoning engines of Gemini, to leveraging spatial intelligence research from Google DeepMind, the collaboration resulted in a system that understands human motion with precision. Unlike traditional motion capture that requires specialised suits, wearable sensors, or multi-camera laboratory setups, the new AI tool turns a standard smartphone into a high-precision sensor, with features including: on-mountain analysis, no wearable gear required, conversational insights and long-term improvements.
“Our collaboration with U.S. Ski & Snowboard is the blueprint for a global shift in how humans move, train and recover, moving beyond historical data to provide athletes with near real-time, prescriptive coaching,” said Oliver Parker, vice president, Global Generative AI, Google Cloud. “By using our full-stack AI, we’re helping democratise elite coaching – proving that if we can solve for the world’s best athletes in the most extreme conditions, we can help anyone from a physical therapy patient to an amateur golfer improve their games.”
U.S. Ski & Snowboard freeski and snowboard athletes and coaches are continuing to prototype this AI tool ahead of the Olympic Winter Games.
For more news like this, click here.