Advancing AI-Driven Sports Intelligence: Dr. Patrick Lucey to Present at NVIDIA GTC 2026

Stats Perform's Chief Scientist will explore how foundational models are transforming broadcast video into scalable, domain-specific insight across soccer and beyond - uncovering and explaining the language of sport through multimodal AI

Erin Lent

~ 4 min read

Dr. Patrick Lucey, Chief Scientist, will take the stage at NVIDIA GTC 2026 in San Jose.

Scale Sports Intelligence and Insight With Foundational Models

MONDAY 16 March | 3:00-3:40 p.m. PDT

Patrick will be joined by Steve Xeller, Chief Revenue Officer, Niall Hendry, VP, AI Applications & Solutions, and Ysabel González-Rico, Team Lead, Core GenAI. The team will be available throughout the event to discuss applied AI strategy, enterprise partnerships and next-generation sports intelligence solutions.

Book a meeting at nvidia gtc

Sport is a complex, multi-agent environment where movement, space and decision-making intersect in real time.

Traditional approaches to sports data rely on fragmented signals or infrastructure-heavy tracking systems. By applying domain-specific foundation models directly to broadcast video, Stats Perform is helping the industry move toward intelligent, scalable sports insight.

The technology combines computer vision and multimodal AI to generate richer representations of match action, supporting generative tracking, tactical modelling and predictive analysis.

Trained on 7.2 petabytes of sports data and supported by more than 575 AI patents, these models are designed specifically for the language of sport.

Inside our session at NVIDIA GTC

In his presentation, Scale Sports Intelligence and Insight With Foundational Models, Patrick will showcase how domain-specific multimodal foundation models are enabling the next generation of sports insight.

The 40-minute session, taking place on 16 March at 3:00 p.m. PDT, focuses on practical innovation rather than theory, exploring how broadcast video can be transformed into structured intelligence that supports performance analysis, content production and fan engagement.

Key themes include:

Modelling match action as continuous spatial-temporal sequences
Enhancing tracking data using generative AI techniques
Developing soccer-specific semantic embeddings for insight discovery
Designing AI systems that integrate directly into product and analytics workflows

The session illustrates how foundation model research is moving beyond theory into real-world sports applications, powering a richer understanding of the game while enabling organisations to scale insight generation globally.

As part of NVIDIA GTC's Sports AI programme, Dr. Lucey joins a line-up that includes leaders from Formula 1, NASCAR and La Liga.

Learn more about the session Add to your calendar

Where sports AI research meets real-world deployment

The foundation models discussed in Patrick's session are not theoretical exercises. They power applied AI solutions already deployed across global sport - supporting performance analysis, content creation and fan engagement at scale.

Stats Perform and NVIDIA are collaborating closely on technical and product solutions to shape the role of AI in the sports market today and define a vision for AI-powered sports technology and fan engagement experiences of the future.

For organisations building next-generation sports platforms, the opportunity lies in how this intelligence integrates into real workflows and commercial environments - transforming advanced AI research into scalable tools for teams, media companies and sportsbooks worldwide.

Meet the Stats Perform team at NVIDIA GTC

Our AI and commercial leadership team will be on-site in San Jose throughout the event and available for scheduled discussions.

Whether you’re exploring AI infrastructure deployment, enterprise partnerships or applied sports intelligence, our team can explain how foundation model architecture translates into scalable, reliable sports-first experiences.