Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Team USA skiers and snowboarders go home with some new equipment, including some gold medals, from 2026 Olympic Games. Besides the years of hard work that went into becoming an Olympic athlete, this year’s crew has an added advantage in their training thanks to a custom AI tool from Google Cloud.
US Ski and Snowboard, the governing body for the U.S. national teams, oversees the training of the country’s best skiers and snowboarders to prepare them for major events, such as the national championships and the Olympic Games. Organization a partner Using Google Cloud to create an AI tool to provide more information about how athletes train and perform on the slopes.
Video review is a big part of winter sports training. The coach will actually stand on the bench to record the athlete’s running, and then review the footage with him afterward to catch mistakes. But the process is somewhat outdated, Anouk Paty, head of sport at US Ski and Snowboard, told me. This is where Google comes in, bringing new AI-powered data insights to the training process.
Google Cloud engineers take to the slopes with skiers and snowboarders to understand how to create an AI model useful for sports training. They used video footage as the basis for the currently unnamed AI tool. Gemini analyzed the video frame by frame, which was then fed into Google DeepMind’s spatial intelligence models. These models were able to take the 2D view of the athlete from the video and turn it into a 3D skeleton of the athlete as they twisted and twisted while running.
An AI model running on screen in the background shows how the tool tracks an athlete’s performance.
Gemini’s finishing touches help the AI tool analyze physics down to pixels, according to Ravi Rajamani, global head of Google’s AI Blackbelt team. Who worked on the project. Coaches and athletes told engineers the specific metrics they wanted to track — speed, spin, and trajectory — and Google engineers coded the model to make it easier for them to monitor and compare different videos. There is also a chat interface to ask Gemini questions about performance.
“From just a video, we are actually able to recreate it in 3D, so you don’t need expensive equipment, (like) sensors, that hinder the athlete’s performance,” Rajamani said.
There’s no denying that instructors are the experts on the mountain, but AI can serve as a kind of gut check. Data can help confirm or disconfirm what coaches see and give them additional insight into the details of each athlete’s performance. It can capture things that humans would have difficulty seeing with the naked eye or with poor video quality, such as where an athlete was looking while performing a trick and the exact speed and angle of the turn.
“It’s data they wouldn’t have gotten otherwise,” Patty said. The 3D skeleton is particularly useful because it makes it easier to see movement otherwise obscured by the puffy jackets and pants athletes wear, she said.
For elite skiing and snowboarding athletes, making small adjustments can mean the difference between a gold medal and no medal at all. Technological advances in training aim to help athletes have all the tools available to improve.
“You’re always trying to find that 1% that can make a difference for an athlete to get them to the podium or win,” Patty said. It can also democratize training. “It’s a way for every coach who works at a club with young athletes to get that level of understanding of what an athlete should do like national team players.”
For Google, a purpose-built AI tool is “the tip of the iceberg,” Rajamani said. There are plenty of potential future use cases, including expanding the basic model to customize it for other sports. It also lays the foundation for work in sports medicine, physical therapy, robotics and ergonomics – specialties where understanding body posture is important. But for now, there’s some satisfaction in knowing that AI was designed to help real athletes.
“This wasn’t a case of technology engineers building something in a lab and delivering it,” Rajamani said. “This is a real problem we’re solving. For us, the motivation was to build a tool that provides a real competitive advantage for our athletes.”