Shield AI, Waabi and GM to Headline Real World AI Stage at TechCrunch Disrupt 2026
Three companies that ship AI into physically unforgiving environments will share the Disrupt 2026 stage to explain how they build systems that cannot afford to fail.
TechCrunch Disrupt 2026 will put three of the hardest cases in applied artificial intelligence on one stage. Executives from Shield AI, Waabi and General Motors are booked for the Real World AI Stage, where the session will center on what it takes to build and deploy AI when a mistake carries physical consequences rather than a bad recommendation. The conference runs in 2026, with early pricing available until September 25 at 11:59 p.m. PT.
The three companies arrive from different corners of the same problem. Shield AI builds autonomous flight software and aircraft for defense missions, where jamming, contested airspace and lost communications are expected conditions. Waabi develops autonomous trucking technology, a domain where perception errors at highway speed are measured in lives. General Motors has spent years pushing driver-assistance and automated driving features into consumer vehicles at production scale, giving it a view of the gap between a working demo and a fleet of millions.
What ties the panel together is the engineering posture each company has adopted. Defense, freight and passenger vehicles all force teams to design for edge cases first, treat simulation and validation as core infrastructure, and accept that the system must degrade safely when its sensors or models are wrong. That is a different discipline from the benchmark-chasing that dominates much of the AI industry, and it tends to shape hiring, testing pipelines and hardware choices.
The choice of speakers also signals where TechCrunch sees the center of gravity in applied AI moving. Consumer chatbots and coding assistants get most of the attention, but autonomy in aircraft, trucks and cars represents a category where regulators, insurers and customers all demand evidence before deployment. The Real World AI Stage is built around that tension between shipping fast and proving safety.
Disrupt has been running similar sessions as autonomy companies move from research milestones toward commercial operations. Waabi has been expanding its trucking program, Shield AI continues to field autonomous aircraft for military customers, and GM has weathered both progress and pullback in its automated driving efforts. Each brings a public track record that will make the discussion concrete rather than theoretical.
The commercial pitch attached to the announcement is straightforward. Early registration pricing ends September 25 at 11:59 p.m. PT, with organizers advertising savings of up to $200, and a second pass available at half price for attendees who bring a colleague. Those deadlines have historically driven the bulk of Disrupt registrations.
The open question for the session is whether the safety-first playbook these companies use can scale beyond high-stakes niches. If validation-heavy development proves too slow or too expensive, the broader AI market may keep choosing speed. If it proves necessary, the rest of the industry could find itself adopting the habits of defense and trucking.