Disrupt 2026 is making artificial intelligence an even bigger focus this year, expanding its singular AI stage into two distinct programming tracks. Alongside the returning AI Stage, the event will debut a brand-new Real World AI Stage that zeroes in on the convergence of digital intelligence and physical reality. The new stage is designed for founders, engineers, investors, and operators who want to understand how AI is moving beyond screens and cloud APIs into vehicles, factories, hospitals, homes, and even biological systems.
The Real World AI Stage will run from October 13 to 15 at Moscone West in San Francisco, in parallel with Disrupt's other stages. The event is expected to bring together more than 10,000 startup and technology leaders, with the new stage hosting some of the most consequential conversations around embodied AI, hardware-software integration, and high-stakes deployment.
Why AI Needs a Real World Track
At past Disrupt events, AI was mostly discussed in the context of software, models, and developer platforms. That is changing quickly. As large language models mature and foundation models become more reliable, a growing wave of startups and tech giants are attempting to embed AI into physical systems that interact with people in real-time. The Real World AI Stage was created to capture this moment, offering a deep dive into autonomous systems that are beginning to operate in public spaces, on battlefields, inside industrial facilities, and within natural ecosystems.
The decision to split the AI stage into two reflects a broader industry realization: AI's future is not just in chatbots or coding assistants, but in machines and biological technologies that can sense, reason, and act in the real world. The programming reflects that, spanning topics from robotic data infrastructure to de-extinction and AI-powered biotech.
Robots Are Waiting for Their ChatGPT Moment
One of the most anticipated sessions on the Real World AI Stage is a conversation led by Les Karpas, Head of Physical AI at Nvidia. Karpas will explore a critical issue that many robotics founders face: the absence of a massive internet-scale dataset for physical intelligence. Large language models had the entire web to train on, and self-driving cars have amassed millions of hours of road data. Robots, by contrast, still lack an equivalent dataset to learn from. This gap is one of the main reasons general-purpose robotic intelligence remains years away, even as algorithmic breakthroughs accelerate across other AI domains.
Karpas will discuss how Nvidia and a new generation of startups are tackling this problem through simulation environments, synthetic data generation, and foundation models tailored for physical AI. The session asks a pivotal question: How close are we to a ChatGPT moment for robots, and what will it actually take to trigger that capability explosion?
Safely Building AI Systems in the Physical World
When AI operates in software, a mistake can be corrected with a patch. But when AI controls an aircraft or an autonomous vehicle, the consequences of failure are far more serious. A session titled “Building AI Systems When Failure Is Not an Option” brings together leaders who are engineering mission-critical AI. Nate Michael, CTO of Shield AI, will share lessons from building autonomous defense technology and industrial systems. Shield AI is known for developing AI pilots for aircraft, so the discussion will center on how founders can create a safety culture, validate AI models for unpredictable environments, and navigate regulatory frameworks in high-stakes industries.
The session will also address trust-building with customers, regulators, and the public, a factor often overlooked by early-stage founders. Michael's perspective is especially valuable because Shield AI's technology operates on battlefields where communication links can be constrained and where the cost of an error is not just financial but potentially human.
Engineering Nature's Comeback With AI
In a conversation that bridges biotechnology and artificial intelligence, Ben Lamm, CEO and co-founder of Colossal Biosciences, will discuss the company's ambitious attempts to bring extinct species back to life. Colossal has attracted significant funding and controversy by applying genetic engineering and AI-driven computational biology to species like the woolly mammoth, the dodo, and the thylacine. Lamm will cover how AI helps analyze complex genomic data, predict gene-editing outcomes, and accelerate the work of synthetic biologists.
The fireside chat will also dive into the ethical and ecological debates surrounding de-extinction. Supporters see it as a breakthrough for conservation, while critics argue it distracts from protecting existing species and ecosystems. Lamm will have the chance to respond directly to those tensions and explain how his company thinks about its role in shaping the future of biodiversity.
AI at the Edge When the Cloud Isn't Available
The real world is not always connected to a data center. In many of the most important AI use cases, cloud connectivity is sparse, latency is unacceptable, or connectivity is deliberately blocked. A session titled “Operating at the Edge: How AI Works When the Cloud Doesn’t” will bring together three experts who have built AI systems for these demanding environments.
Dr. Ali Agha, CEO and founder of FieldAI, works at the intersection of space exploration and field robotics. Michelle Lee, CEO and founder of Medra, builds AI for medical diagnostics in settings that may not have reliable internet or centralized processing. Aidan Madigan-Curtis, partner at Eclipse Ventures, will contribute an investor's view on which edge-AI startups are poised to scale. The panel will discuss architectural principles, design trade-offs, and real-world lessons for making AI work when offline is the default.
From Prototype to Production in Deep Tech
Many deep tech startups struggle with a specific transition: taking a working prototype and turning it into a manufactured product that can scale profitably. A prototype that works in a lab is not a product; a product that ships is not a scaled business. Most failures happen in the gap between those stages. This session brings together founders who have crossed that gap in space hardware, humanoid robotics, and autonomous systems.
John Mackey, CEO and co-founder of MBRYONICS, will share lessons from aerospace and biotech hardware. Boris Sofman, co-founder and CEO of Bedrock Robotics, will talk about humanoid robotics manufacturing. Adrian Macneil, CEO of Foxglove, brings a developer perspective from building tools that robotics and autonomous vehicle teams use to debug real-world deployments. Together, they will cover what they got wrong, what they should have done earlier, and how supply chains and manufacturing realities shape engineering decisions.
More Than Just a Stage
The Real World AI Stage is only one part of a much larger event. Disrupt 2026 will also feature the Disrupt Stage, which hosts top speakers from Replit, Amazon, Tether, and other major companies. The AI Stage will offer conversations around security gaps and business model shifts affecting software companies. The Smart Money Stage will deal with stablecoins, instant payments, and AI's role in financial trust. The Smart Systems Stage will explore fusion breakthroughs and grid strain. The Builders Stage will provide tactical advice for founders and investors on raising capital, hiring, and scaling.
In total, Disrupt 2026 promises six industry-specific stages, a Startup Battlefield competition, an exhibition floor with hundreds of startups and companies, and countless networking opportunities. Attendees can register at discounted rates by taking advantage of the promotion mentioned during the announcement, which offers up to $300 off standard ticket pricing.
The addition of the Real World AI Stage signals where the technology industry is headed. As AI becomes more tangible, operating in everything from defense systems to laboratory biology, the need for practical conversations at the intersection of code and physical reality grows. Disrupt 2026 is positioning itself as the venue where those conversations happen, and the lineup on the Real World AI Stage is likely to draw some of the most engaged audiences in San Francisco this October.
Source: TechCrunch News