TechCrunch Disrupt 2026 will introduce a new Real World AI Stage, examining how artificial intelligence is moving beyond software and into robots, autonomous vehicles, defence systems, industrial machinery and biology.
The additional programme will run alongside the event’s existing AI Stage at San Francisco’s Moscone West from October 13 to 15. Speakers will include senior figures from Nvidia, Shield AI, Colossal Biosciences, FieldAI, Foxglove and other technology companies.
Real World AI Stage to examine robotics and autonomous systems
One of the central sessions will explore what is needed to give robots a breakthrough comparable to the rapid advances seen in large language models.
Les Karpas, Nvidia’s head of physical AI, will discuss the shortage of real-world data available to robotic systems. While language models have been trained on vast quantities of internet content and autonomous vehicles have accumulated extensive road data, robots have far less information from which to learn.
The session will consider whether new data pipelines, simulation environments and foundation models can close that gap, and how close the industry may be to broadly capable robotic intelligence.
Safety will be another major theme, with Nate Michael, chief technology officer at defence technology company Shield AI, due to examine the difficulties of deploying autonomous systems when errors can have serious consequences.
The discussion will cover testing, validation, regulation and the creation of safety cultures within companies developing autonomous vehicles, defence technologies and industrial systems.
Colossal chief to discuss de-extinction technology
Ben Lamm, chief executive of Colossal Biosciences, will take part in a fireside conversation about the company’s work on de-extinction and the use of artificial intelligence in modern biology.
The session will address the technologies being developed to revive extinct species, as well as the wider debate over whether engineering nature could become a conservation tool or divert attention from protecting species that still exist.
Other sessions will focus on the technical challenges of running AI systems away from the cloud, where internet connectivity may be limited and rapid responses are essential.
Dr Ali Agha, founder and chief executive of FieldAI, Medra founder and chief executive Michelle Lee, and Aidan Madigan-Curtis, a partner at Eclipse Ventures, will discuss the design choices involved in deploying AI in defence, space and industrial environments.
The panel is expected to examine latency, connectivity and reliability, along with the trade-offs involved in building systems that must operate at the edge rather than relying on remote data centres.
Founders to discuss taking deep-tech ideas into production
A further session will look at the difficult journey from a working prototype to a commercially viable product, an area where many deep-tech businesses struggle as manufacturing and supply-chain demands replace laboratory conditions.
John Mackey, co-founder and chief executive of MBRYONICS, Boris Sofman, co-founder and chief executive of Bedrock Robotics, and Adrian Macneil, chief executive of Foxglove, will share lessons from developing space hardware, humanoid robotics and autonomous systems.
The programme will sit within a wider TechCrunch Disrupt 2026 event featuring stages devoted to startups, finance, smart systems and the broader technology industry.
The Disrupt Stage will include speakers from Replit, Amazon and Tether, while the Smart Money Stage will cover stablecoins, instant payments and the role of AI in financial trust.
The Smart Systems Stage will examine advances in fusion and the pressure that growing AI demand is placing on energy grids. A Builders Stage will focus on practical issues including fundraising, recruitment and scaling a company.
Organisers said more than 10,000 startup, technology and venture capital leaders are expected to attend the three-day event, which will also feature the Startup Battlefield and an exhibition floor.
