Shield AI, Waabi, and General Motors on building AI when failure is not an option at TechCrunch Disrupt 2026

San Francisco is buzzing with anticipation as TechCrunch Disrupt 2026, a premier event for startups and tech innovators, prepares to open its doors from October 13-15 at Moscone West. This year, a significant highlight is the newly introduced "Real World AI Stage," which promises deep dives into the practical, high-stakes applications of artificial intelligence. Among the most anticipated sessions is a panel discussion featuring leaders from Shield AI, Waabi, and General Motors, who will tackle the critical topic of "building AI when failure is not an option."

The conversation is set to bring together Nathan Michael, Chief Technology Officer at Shield AI; Raquel Urtasun, Founder and CEO of Waabi; and Mikell Taylor, Director of Robotics Strategy at General Motors. These industry pioneers will share their unique perspectives on developing and deploying AI systems in environments where precision, reliability, and safety are paramount, and the consequences of error extend beyond a screen to the physical world.

The Imperative of Flawless AI in Critical Applications

As AI continues its rapid evolution, moving from theoretical models and digital interfaces into tangible machines that interact with our physical environment, the discussion around safety and reliability intensifies. The "Real World AI Stage" at TechCrunch Disrupt 2026 is specifically designed to address this intersection of digital and physical AI.

For companies like Shield AI, Waabi, and General Motors, the stakes couldn't be higher. Their AI systems are not just recommending products or generating text; they are piloting aircraft, driving heavy trucks, and controlling complex manufacturing processes. In these scenarios, a single misstep can lead to catastrophic outcomes, making the concept of "failure is not an option" a fundamental guiding principle. The upcoming session will delve into how these organizations approach testing, validation, regulatory hurdles, and fostering a safety-first culture in their pursuit of robust AI.

Shield AI: Autonomous Defense and the Battlefield

Shield AI, founded in 2015, is an American aerospace and defense technology company based in San Diego, California. Their core mission is to develop AI-powered fighter pilots, drones, and advanced technology for defense operations, aiming to shift dangerous tasks from human service members to autonomous systems. The company's clients include the United States Special Operations Command, US Air Force, US Marine Corps, US Navy, and several international militaries.

One of Shield AI's notable achievements is its Nova autonomous quadcopter drone, which was the first AI-powered drone to be deployed for defense purposes in US military history. They also developed Hivemind, an autonomy and AI stack that enables drones and aircraft to maneuver autonomously even in GPS-denied or communication-degraded environments. This technology is crucial for reconnaissance in close-quarters combat and solving complex problems like room-clearing.

The importance of "failure is not an option" is profoundly evident in defense applications. In combat scenarios, the reliability of autonomous systems directly impacts human lives and national security. Shield AI's CTO, Nathan Michael, will likely share insights into their rigorous development processes, extensive testing protocols, and the ethical considerations involved in deploying AI in such critical contexts. The company has raised over $1.3 billion in capital and was valued at approximately $5.3 billion in March 2025, reflecting significant investor confidence in its cutting-edge defense autonomy solutions.

Waabi: Revolutionizing Autonomous Trucking with AI

Waabi Innovations Inc., headquartered in Toronto, Canada, is an artificial intelligence company founded in 2021 by Raquel Urtasun, a former chief scientist and head of R&D at Uber's Advanced Technologies Group. Waabi focuses on developing self-driving technology, particularly for commercial self-driving trucks and next-generation robotaxi applications.

The company's approach centers on a simulation-first, generative AI-heavy development methodology, aiming to achieve commercial scale for autonomous trucks faster and more cost-effectively than traditional mile-accumulation methods. Their flagship product, the Waabi Driver, is an end-to-end AI system designed to make decisions and adapt to new scenarios and hardware configurations. A key component of their technology is Waabi World, a realistic and scalable simulation environment launched in 2022, which allows for extensive testing and learning.

Autonomous trucking presents immense challenges where failure is simply unacceptable. A malfunction in a 40-ton vehicle on a highway can have devastating consequences. Waabi's commitment to safety is underscored by its focus on neural simulation and end-to-end AI to set new safety standards. In January 2026, Waabi secured a significant $750 million Series C funding round, with participation from Nvidia, Volvo Group, and Porsche, pushing its total funding to over $1 billion. Waabi has also partnered with Uber Freight and Volvo Autonomous Solutions for commercial deliveries and joint development of self-driving technology in long-haul trucks.

Raquel Urtasun's participation in the TechCrunch Disrupt panel will offer invaluable insights into how Waabi is tackling the complexities of real-world autonomous vehicle deployment, from robust AI training in simulation to navigating regulatory landscapes and building public trust.

General Motors: AI Across the Automotive Ecosystem

General Motors (GM), a global automotive giant, is deeply integrating AI across its operations, from manufacturing to in-car experiences and autonomous driving. The company is actively leveraging AI to reimagine how vehicles are designed, built, supported, and experienced, moving beyond traditional mechanical engineering excellence.

GM's AI initiatives include predictive manufacturing, where AI-driven analytics optimize vehicle output by analyzing historical performance and real-time data from robotics and conveyor systems. This helps in early detection of potential malfunctions, preventing unscheduled downtimes. They also use AI and digital twins to simulate production lines before construction, optimizing planning and accelerating scaling.

In the realm of in-car technology, GM has partnered with Google Cloud to integrate conversational AI into vehicles through the OnStar Virtual Assistant. This system uses advanced intent-recognition algorithms to provide quicker assistance and allow human advisors to focus on emergencies. Furthermore, GM is developing a new vehicle AI assistant, launched later in 2026, that combines conversational AI with vehicle-specific data and OnStar intelligence for a deeper, more personalized ownership experience.

While autonomous driving is a significant area where "failure is not an option" applies, GM's AI Director of Robotics Strategy, Mikell Taylor, will likely discuss the broader implications of AI safety across the entire automotive lifecycle. This includes AI-powered design and aerodynamics, where AI tools help engineers make better decisions earlier in vehicle development. The insights from GM will highlight the scale at which a legacy automaker is embedding AI into safety-critical functions and the comprehensive strategies required to ensure reliability in mass-produced vehicles.

The Broader Significance: Trust, Regulation, and the Future of AI

The collective insights from Shield AI, Waabi, and General Motors at TechCrunch Disrupt 2026 will underscore a critical turning point in AI development. As AI systems become increasingly integrated into the physical world, the conversation shifts from mere technological capability to absolute trustworthiness. The session, "Building AI Systems When Failure Is Not an Option," directly addresses the core concerns of safety, reliability, and ethical deployment that are central to the future acceptance and success of AI.

Discussions will likely cover the importance of diverse datasets, advanced simulation techniques, robust validation frameworks, and clear regulatory guidelines. Building a safety culture within AI development teams, establishing transparent accountability mechanisms, and fostering public understanding and trust are all essential components when the consequences of AI failure are physical and potentially life-threatening. This gathering of leaders from defense, autonomous logistics, and mainstream automotive industries highlights a shared commitment to pushing the boundaries of AI responsibly. Their experiences will offer a roadmap for other deep-tech founders and innovators navigating the complex landscape of real-world AI deployment.

Conclusion

The TechCrunch Disrupt 2026 Real World AI Stage is set to be a pivotal forum for addressing the most challenging aspects of AI deployment. The session featuring Shield AI, Waabi, and General Motors will not just discuss technological advancements but will emphasize the profound responsibility that comes with building AI systems where failure carries unacceptable risks. Their shared experiences and strategies will undoubtedly shape the discourse around safety-critical AI, pushing the industry towards more robust, reliable, and trustworthy autonomous solutions for the future. The opportunity to learn from these leaders on October 13-15 in San Francisco is a must for anyone invested in the responsible evolution of artificial intelligence.

Frequently Asked Questions

What is the "Real World AI Stage" at TechCrunch Disrupt 2026?

The Real World AI Stage is a new dedicated stage at TechCrunch Disrupt 2026, running from October 13-15 in San Francisco. It focuses on the intersection of digital and physical AI, covering topics like robotics, autonomous transportation, defense systems, and biotechnology, emphasizing how AI is being embedded into tangible machines that interact with our world.

Who are the key speakers on the "Building AI Systems When Failure Is Not an Option" panel?

The panel will feature Nathan Michael, Chief Technology Officer at Shield AI; Raquel Urtasun, Founder and CEO of Waabi; and Mikell Taylor, Director of Robotics Strategy at General Motors.

What makes building AI for these companies particularly challenging?

These companies operate in domains where AI errors can have severe physical consequences, affecting human lives, safety, and critical infrastructure. Shield AI deals with defense applications, Waabi with autonomous trucking, and General Motors with self-driving cars and advanced manufacturing. The high stakes demand exceptionally robust, reliable, and rigorously tested AI systems.

What specific topics will the panel cover regarding AI safety?

The discussion will delve into creating a strong safety culture within AI development, the methodologies for rigorous testing and validation of AI systems, navigating complex regulatory barriers, and building companies that can earn and maintain trust when the consequences of failure are physical and significant.