Agenda
The Worlds AI Needs: Games as Infrastructure for Intelligence
Matthew Lee (Lead Generative AI Analyst, NVIDIA)
Pass Type:
Academic Pass, Festival Pass, Game Changer Pass, Indie Pass, Start-Up Pass -
Get your pass now!
Hot Topics:
AI
Track:
Luminaries Speaker Series
Audience Level:
Advanced
Format:
Lecture
Vault Recording: TBD
Audience Level: Advanced
The games industry has long asked, "What will AI do to games?" Fundamentally, the industry's relationship with AI has been as consumers and users, exploring how AI can change the way games are made and whether it can make them more engaging.
What the industry should be asking is: "What can games do for AI?"
As AI shifts from generative systems toward models that are agentic, multimodal, and capable of acting in the world, it increasingly needs environments where intelligence can perceive, decide, act, experience consequences, and learn to operate over time. Games have spent decades building exactly those kinds of environments for intelligent actors - that is, for players.
From platformers teaching simple mechanics to roguelikes where actors learn across iterations, multiplayer games enabling cooperation and competition, and persistent worlds with evolving history and consequences that endure, games—and the people who build them—have a great deal to offer.
This session explores the shifting relationship between games and AI, examining the strategic implications for game companies as the demands of AI development change. It considers the worlds, tools, data, and expertise the industry has developed, the emerging opportunities some companies are already pursuing, and the challenges that may limit or shape those opportunities.
And in a period of industry contraction and market upheaval, this session argues that a company's most strategic assets may not be its existing products and IP, but its capabilities, personnel, and institutional knowledge—the ability to create the worlds of tomorrow.
What the industry should be asking is: "What can games do for AI?"
As AI shifts from generative systems toward models that are agentic, multimodal, and capable of acting in the world, it increasingly needs environments where intelligence can perceive, decide, act, experience consequences, and learn to operate over time. Games have spent decades building exactly those kinds of environments for intelligent actors - that is, for players.
From platformers teaching simple mechanics to roguelikes where actors learn across iterations, multiplayer games enabling cooperation and competition, and persistent worlds with evolving history and consequences that endure, games—and the people who build them—have a great deal to offer.
This session explores the shifting relationship between games and AI, examining the strategic implications for game companies as the demands of AI development change. It considers the worlds, tools, data, and expertise the industry has developed, the emerging opportunities some companies are already pursuing, and the challenges that may limit or shape those opportunities.
And in a period of industry contraction and market upheaval, this session argues that a company's most strategic assets may not be its existing products and IP, but its capabilities, personnel, and institutional knowledge—the ability to create the worlds of tomorrow.
Takeaway
Attendees learn how games can support physical, agentic, and long-horizon AI development, with new opportunities emerging from gameplay-derived training data, virtual worlds, and persistent environments. They gain a framework for identifying which existing assets and capabilities may be valuable, where adaptation is required, and which opportunities are worth pursuing.