TechCrunch Disrupt 2026 is doubling down on physical AI. The conference, running October 13-15 at San Francisco’s Moscone West, is splitting its AI programming into two stages: the traditional AI Stage and a brand-new Real World AI Stage focused on the messy intersection of digital intelligence and physical hardware.
The new stage’s lineup reads like a who’s-who of companies betting that AI’s next frontier isn’t a chatbot — it’s a robot, a drone, or a resurrected species. Nvidia’s Les Karpas, head of physical AI, will tackle the data gap that’s keeping general-purpose robots from having their own ChatGPT moment. Shield AI CTO Nate Michael will discuss what safety culture looks like when a software bug can crash a vehicle or compromise a military mission. And Colossal Biosciences CEO Ben Lamm will defend his controversial billion-dollar bet on de-extinction.
The most interesting thread running through the programming is a question the AI industry has mostly avoided: what happens when failure isn’t just an awkward hallucination but a physical catastrophe? Edge computing gets its own session with FieldAI’s Dr. Ali Agha and Medra’s Michelle Lee, who build systems where the cloud is unreachable and latency means lives. It’s a sharp contrast to the SaaS-centric AI conversations that have dominated recent years.
What’s notably absent from the lineup is any big consumer robotics company. No Figure, no Tesla Optimus team, no Boston Dynamics. Instead, the stage skews heavily toward defense, industrial, and scientific applications — a signal that the most credible near-term deployments of physical AI are happening where budgets are large, stakes are high, and public scrutiny is lower. That’s probably the right call, though it does raise the question of whether consumer robotics will ever get its iPhone moment, or if it’ll remain a perpetual demo reel.
