At the upcoming TechCrunch Disrupt 2026, the focus on artificial intelligence is expanding with the debut of a new 'Real World AI Stage'. This dedicated platform will explore the crucial intersection between digital advancements and their physical manifestations, examining how AI is increasingly integrated into our daily lives, from autonomous systems in public spaces to revolutionary breakthroughs in biotechnology. The event, scheduled from October 13 to 15 in San Francisco, promises to bring together innovators and leaders who are shaping the future of AI in tangible, impactful ways, featuring discussions that span from the challenges of robot development to the ethical debates surrounding de-extinction.
This expanded focus highlights the rapid evolution of AI, moving beyond theoretical models to practical, real-world applications. The stage will address the complexities of deploying AI in critical environments where precision and reliability are paramount, and explore the technological hurdles and solutions in developing AI for scenarios where traditional cloud-based systems are insufficient. Attendees will gain valuable insights into the architectural principles, design choices, and manufacturing realities essential for transforming AI prototypes into scalable, market-ready products, thus offering a comprehensive look at the ecosystem driving the next generation of intelligent systems.
The Road to General Robotic Intelligence: Bridging the Data Gap
The advancement of general-purpose robotic intelligence faces a significant hurdle: the lack of extensive, real-world data comparable to what large language models and self-driving cars have access to. This data scarcity is a primary reason why widespread autonomous robotic capabilities remain a future aspiration. However, a new generation of startups is actively working to overcome this challenge by developing innovative data pipelines, sophisticated simulation environments, and foundational models, aiming to catalyze a similar surge in capabilities for physical AI as seen with large language models.
This session will critically examine what is truly required for physical AI to experience its 'ChatGPT moment' and assess how close we are to achieving it. Les Karpas, Head of Physical AI at Nvidia, will lead discussions on the technological and data infrastructure necessary to bridge the current gap. The focus will be on understanding the foundational requirements and the ongoing efforts to create robust data ecosystems that can support the complex learning and adaptation needs of advanced robots, pushing them closer to versatile, real-world functionality.
Ensuring AI Reliability in High-Stakes Environments
When artificial intelligence is deployed in physical environments, the repercussions of system failure can be severe, ranging from industrial accidents to critical operational breakdowns. This reality underscores the paramount importance of developing AI systems with inherent safety and reliability, especially in sectors like autonomous vehicles, defense technology, and industrial automation. Leaders in these fields are continually grappling with the critical question of how to determine when an AI system is sufficiently robust and safe for real-world deployment.
Nate Michael, CTO of Shield AI, will contribute to this discussion, sharing insights into fostering a strong safety culture within AI development. The session will explore essential practices for rigorously testing and validating AI systems, navigating complex regulatory landscapes, and building companies that can earn and maintain trust in high-stakes applications. Practical strategies for ensuring AI systems operate flawlessly where errors are not an option will be highlighted, offering valuable lessons for founders and engineers in hard tech.
