X Games taps Google AI to judge snowboarders

In extreme snow sports, a tiny hesitation on a landing or a half-inch drift in the air can separate gold from a forgettable run. For decades, those split-second margins rested entirely on human eyes braving freezing temperatures. Now, Google is stepping into the judges' booth.

X Games announced a partnership with Google Cloud to test an experimental AI system during X Games Aspen 2025. Running from January 23 to January 25, the sold-out event will put the tech through its paces during the men's and women's Superpipe competitions.

The system is built on Google's Vertex AI and large language models, trained directly on competition footage to understand tricks, landing qualities, and motion economy. One of the lead human judges from X Games helped build it, while athletes noted the system feels like having a coach in your pocket.

Predicting podiums and calling the action

The AI does more than just tally points. During early practice sessions, officials will feed it practice runs and ask it to predict the top three finishers among snowboarders.

Once live runs begin, the model breaks down video feeds to extract hard metrics, including airtime, trick difficulty, and landing execution. It is also wired to deliver play-by-play commentary across an entire run, weaving in athlete history and broadcasting in multiple languages, such as Japanese.

X Games CEO Jeremy Bloom, a two-decade veteran athlete, noted that snow sports have long lacked the instant replay tools used in basketball. With modern computer vision, he argued, basic judging errors simply shouldn't happen anymore.

Human judges keep the medals

Despite the big claims, the AI won't be handing out any actual X Games medals this weekend. Bloom made it clear that the technology is still experimental and far from perfect.

Instead, human judges will determine all official podium finishes. The AI score will simply be broadcast to the live crowd on a slight time delay, allowing fans to compare the machine's verdict directly against human scores.

The trial also sets the stage for broader format shifts. By 2026, X Games plans to launch the Action Sports Games League (XGL), where real-time AI performance analytics will feed into fantasy sports, betting, and deeper fan engagement.

Why it matters

Subjectively scored sports have always walked a fine line between technical perfection and stylistic expression. Disagreements over bias and consistency are practically a tradition. Bringing in video-based LLMs provides an objective baseline for parameters humans struggle to measure precisely, like exact hang time and landing angles.

For fans, this turns high-speed acrobatics into understandable data. Instead of guessing why one run scored higher than another, viewers get instant breakdowns and multilingual coverage that lowers the barrier to entry.

We think X Games played this smart by keeping the AI strictly in an advisory and exhibition role. Generative models and computer vision can still misinterpret novel tricks or creative flair that athletes invent on the fly. By running the system as an open benchmark alongside human judges rather than replacing them, X Games gets all the engagement perks of new tech without risking the integrity of the podium.