TechCrunch Disrupt 2026 launches Real World AI Stage
TechCrunch Disrupt 2026 will launch a new Real World AI Stage alongside its existing AI Stage, broadening the event’s AI programming to focus on systems that intersect with the physical world. The event is scheduled for October 13–15 at Moscone West in San Francisco and is expected to bring together more than 10,000 startup, technology and venture-capital leaders.
The new stage will examine autonomous hardware and AI deployments beyond self-driving cars, including systems used in public spaces, defence, homes, industrial settings and biology. Its announced lineup includes speakers from Nvidia, Shield AI, Colossal Biosciences, FieldAI and Foxglove, with further participants still to be announced.
From robotics data to safety-critical deployment
One session, Robots Are Waiting for Their ChatGPT Moment, will address the data problem facing general-purpose robotics. TechCrunch frames the issue in contrast to large language models, which drew on internet-scale data, and self-driving vehicles, which accumulated millions of hours of road data. Les Karpas, Nvidia’s Head of Physical AI, will discuss the data pipelines, simulation environments and foundation models that startups are building to narrow that gap.
Another session will consider AI systems for which failure carries immediate consequences, such as grounded aircraft, vehicle crashes or compromised missions. Nate Michael, CTO of Shield AI, is set to discuss safety culture, testing and validation, regulatory hurdles and the challenge of earning trust before autonomous vehicles, defence technologies and industrial systems are deployed.
The wider AI agenda also includes AI Stage security and SaaS business shifts examining security gaps and business-model shifts affecting SaaS companies, while the Real World AI Stage concentrates on the engineering demands of AI that interacts with physical environments.
Edge AI, biology and the production gap
Edge computing is another core subject. Dr. Ali Agha of FieldAI, Michelle Lee of Medra and Aidan Madigan-Curtis of Eclipse Ventures will discuss AI deployments in defence, space and industrial settings where latency matters, connectivity is limited and cloud access may not be available. The session is intended to cover architectural principles, design choices and trade-offs for systems operating under those constraints.
Ben Lamm, CEO of Colossal Biosciences, will address the technologies behind efforts to revive extinct species, AI’s role in modern biology and the debate over whether engineering nature supports conservation or distracts from protecting existing species.
A separate production-focused discussion will feature John Mackey of MBRYONICS, Boris Sofman of Bedrock Robotics and Adrian Macneil of Foxglove. It will focus on the transition from a working prototype to production and then to profitable volume, where supply chains and manufacturing realities can challenge deep-tech companies.
For businesses evaluating physical AI, the practical implication is clear: capability demonstrations are only one stage of adoption. Data collection, validation, connectivity, regulation, supply chains and production planning must be assessed together before an AI system can move from a lab or pilot into dependable operations.