TechCrunch Disrupt 2026 will cover AI safety, enterprise applications, and physical systems in five sessions. The event will take place at Moscone West in San Francisco on October 13-15. More than 10,000 founders, investors, executives, and technology leaders are expected to attend the event, which will feature 200+ sessions, 250+ speakers, and 300+ exhibiting startups.
The sessions and topics to be covered are as follows:
- In the “What Anthropic Sees When Enterprises Actually Deploy Claude” session, which will feature Anthropic Head of Applied AI Cat de Jong, the discussion will examine which applications succeed as companies move from AI pilots to production environments and why some remain in the pilot phase even after 18 months.
- In the “The Agent Security Problem Nobody Is Talking About” session, featuring Okta Chief Product and Technology Officer Ric Smith and NanoCo co-founder and CEO Gavriel Cohen, the discussion will cover AI agents’ access permissions, weaknesses in application-level permission models, and infrastructure decisions for agent security.
- AWS Vice President of Security Services Rudy Mitra, Luta Security CEO Katie Moussouris, and cybersecurity expert Wendy Nather will assess the security, governance, observability, and infrastructure requirements for enterprise AI in the “Securing the AI Enterprise: Why the Cloud Just Got a Lot More Complicated” session.
- Shield AI Chief Technology Officer Nathan Michael, General Motors Director of Robotics Strategy Mikell Taylor, and Waabi founder and CEO Raquel Urtasun will discuss when the safety of autonomous systems can be considered sufficient in the “Building AI Systems When Failure Is Not an Option” session. Safety culture, testing and validation processes, and regulatory barriers will also be on the agenda.
- In the “Robots Are Waiting for Their ChatGPT Moment. Here Is What Is Standing in the Way” session, Nvidia Inception Global Head of Physical AI Les Karpas will examine the lack of data in robotics, data pipelines, simulation environments, and foundation models.
Tickets are being offered with discounts of up to $200 through 11:59 p.m. PT on September 25, along with 50% off the second ticket. The event will also discuss AI, security, reliability, and the challenges standing in the way of trust in the real-world use of technologies.
Why it matters
The program shows that AI will be evaluated not only by model performance, but also by how it is embedded in organizational processes and what permissions it is given to operate. This framework puts questions around access permissions, traceability, and governance at the forefront for companies using AI agents, security teams, and infrastructure providers. When autonomous vehicles and robots are involved, the low tolerance for error makes regulatory conditions as decisive as testing and validation methods. The common thread among the discussions is the question of whether technical capacity alone is sufficient for systems to be used reliably in the real world. In particular, how gaps in data, simulation, and application-level security vulnerabilities will be addressed remains among the open questions.
Background
Disrupt is not a new name in the FikirPilot archive: we have published 12 articles mentioning the name in the last 90 days; the most recent one is dated September 24, 2026.
Term: agent
An AI agent is software that calls tools and carries out multi-step tasks to achieve a goal rather than producing a single response.