When Failure Is Not an Option: Industry Leaders to Tackle High-Stakes AI at TechCrunch Disrupt 2026

In the realm of digital assistants and generative software, a malfunctioning AI chatbot might offer a hallucinated fact or a grammatically awkward sentence. While frustrating, the consequences are typically contained within the screen. However, as artificial intelligence transitions from the digital workspace into the physical world—taking the helm of aircraft, autonomous vehicles, and industrial robotics—the stakes shift dramatically. In these high-stakes environments, a software error is not merely a bug; it is a potential catalyst for a mission failure, a vehicle collision, or a compromised operation.

As the industry grapples with the transition from lab-tested prototypes to real-world deployment, a critical question looms over developers and regulators alike: How do you definitively prove an autonomous system is ready to be unleashed into the wild?

At TechCrunch Disrupt 2026, this vital inquiry will take center stage. Three pioneers of the autonomous movement—Shield AI’s Chief Technology Officer Nathan Michael, Waabi founder and CEO Raquel Urtasun, and General Motors’ Director of Robotics Strategy Mikell Taylor—will convene on the Real World AI Stage to explore the complexities of "Building AI Systems When Failure Is Not an Option." The panel promises a deep dive into the engineering and ethical rigor required to move beyond controlled demonstrations toward full-scale, reliable autonomy.

Navigating the High-Stakes Frontier

The upcoming discussion at Disrupt is designed to unpack the multifaceted work required to deploy AI where the consequences of error are absolute. The panelists will address the necessity of cultivating a robust safety culture, the technical hurdles of rigorous system validation, the evolving landscape of regulatory oversight, and, perhaps most importantly, the psychological challenge of earning public and institutional trust.

For those building at the intersection of AI and the physical world, the window to join this conversation is closing. TechCrunch has announced that September 25 at 11:59 p.m. PT serves as the final deadline for attendees to secure a Disrupt pass with a discount of up to $200. With price increases set to follow, this deadline marks a pivotal opportunity for engineers, investors, and stakeholders to gain insights from three leaders who operate daily in environments where there is zero margin for error.

Ready for the Mission: The Defense Perspective

The development of mission-critical autonomy requires a departure from traditional software development cycles. At the forefront of this effort is Nathan Michael of Shield AI. Michael is the architect behind Hivemind, a platform-agnostic mission autonomy software designed to operate in some of the most challenging and unpredictable environments on Earth.

Michael’s career is rooted in a deep understanding of control, perception, and multi-robot systems. His extensive tenure at Carnegie Mellon University’s Robotics Institute, where he directed the Resilient Intelligent Systems Lab, provided the academic foundation for his current work. Today, that expertise is being applied to the defense sector, where the pressure for performance is matched only by the necessity for absolute assurance.

The relevance of Michael’s work was underscored in February when Hivemind was selected as a primary autonomy provider for the U.S. Air Force’s Collaborative Combat Aircraft (CCA) drone prototype program—a testament to the military’s growing reliance on autonomous, AI-driven assets. Shortly thereafter, in March, Shield AI made waves in the investment community, announcing a $1.5 billion Series G funding round that pushed the company’s post-money valuation to a staggering $12.7 billion. At Disrupt, Michael will offer a rare look at the methodologies required to ensure that autonomy can be trusted with national security objectives.

Ready for the Road: The Path to Driverless Deployment

While the defense sector focuses on tactical missions, the autonomous vehicle industry faces the massive challenge of navigating the public sphere. Raquel Urtasun, the founder and CEO of Waabi, has spent over two decades navigating this complexity. With a storied background that includes a tenure as chief scientist and head of R&D at Uber ATG, Urtasun is a veteran of the autonomous vehicle wars. Beyond her role at Waabi, she remains a prominent figure in academia as a professor of computer science at the University of Toronto and a co-founder of the Vector Institute for AI, having contributed to over 200 scholarly papers.

Waabi is currently aiming to redefine the industry’s approach to deployment. In January, the company announced a $1 billion raise and a high-profile partnership with Uber, aimed at the deployment of at least 25,000 autonomous trucks powered by the "Waabi Driver" system.

However, Urtasun has been vocal about the importance of a measured approach. The cornerstone of her company’s strategy is the Waabi World simulator, a virtual environment designed to train and rigorously stress-test the system before it ever touches a public highway. Urtasun maintains that, regardless of the promise shown in simulation, autonomous trucking requires a level of validation that transcends current industry standards. Her insights at Disrupt will be particularly illuminating for those seeking to understand the threshold at which a vehicle can be considered truly ready to operate without human intervention.

Ready to Work Alongside People: The Industrial Robot Perspective

The challenge of autonomy is not limited to vehicles and aircraft; it is also a fundamental hurdle in the industrial sector. Mikell Taylor, who currently leads robotics strategy for General Motors’ Autonomous Robotics Center, has dedicated over twenty years to the field. Her career is marked by a focus on practical, reliable machines that can operate effectively alongside human workforces.

Before her current role at GM, Taylor led the development of Proteus at Amazon Robotics, the company’s first fully autonomous mobile robot. Her journey into robotics, which reportedly began with a project to build a robotic senior prom date, has evolved into a career defined by high-level engineering and a deep focus on user experience.

Taylor’s presence on the Disrupt stage is critical because she brings a focus on human-machine interaction. She argues that for robots to be successfully integrated into workplaces, the human element—user experience, safety, and adoption—must be baked into the design process from the very beginning. When AI-powered machines leave the safety of a controlled lab and enter a dynamic warehouse or manufacturing plant, the ability of the system to communicate its intentions to human coworkers becomes as important as its ability to navigate obstacles.

The Standard of Confidence

Aircraft, autonomous vehicles, and industrial robots operate in vastly different contexts, yet they all face the same fundamental test: the moment of truth where theory meets reality. For founders and technology leaders who are currently building the next generation of autonomous systems, the upcoming session at TechCrunch Disrupt offers a rare, cross-disciplinary opportunity. It allows attendees to compare how leaders in defense, autonomous driving, and industrial robotics manage the critical decisions that separate a successful product from a failed one.

This panel is one of more than 200 sessions scheduled across six specialized stages at Disrupt, which will take place from October 13-15 at Moscone West in San Francisco. The event is set to host over 10,000 attendees, including a robust mix of founders, investors, and operators, along with 250 speakers and more than 300 exhibiting startups. Beyond the scheduled programming, the conference is designed to facilitate the high-level networking, matchmaking, and dealmaking necessary to push the industry forward.

For those committed to the idea that "almost ready" is never enough when the stakes involve human safety and operational success, the clock is ticking. The final opportunity to secure a discounted pass ends on September 25 at 11:59 p.m. PT. As the industry moves toward a future where autonomous systems are increasingly pervasive, understanding the standard of "readiness" has never been more essential. By participating in this conversation, attendees will gain clarity on what it takes to deploy technology that the world can not only rely on, but trust.