TechCrunch Disrupt 2026 has released the full agenda for its breakout sessions, offering a detailed look at the practical challenges facing AI founders and operators this year. The conference, scheduled for October 13-15 at Moscone West in San Francisco, will feature interactive, 50-minute sessions designed to facilitate direct Q&A between attendees and industry leaders.
What Happened
The newly revealed lineup focuses on specific operational and strategic questions currently dominating the AI sector, such as how to build with AI agents, where value accrues as the industry shifts from model training to inference, and the definition of physical AI. Notable sessions include "Outnumbered Either Way," where Anthropic Members of Technical Staff Steve Androulakis and Sachin Malhotra will demonstrate how they run multiple Claude agents in parallel to delegate engineering work and review output. Another key session, "The Inference Economy," features Hagay Lupesko, SVP of AI Cloud at Cerebras Systems, and Rudina Seseri of Glasswing Ventures, who will discuss why defensibility is shifting away from the model itself and toward latency and infrastructure decisions.
Other agenda items address the emerging category of physical AI, with sessions led by investors and builders from SOSV HAX, Venrock, and Archetype AI. These discussions aim to distinguish between perception and action in autonomous systems, exploring the execution layer required for industrial AI applications. Additionally, sessions on fundraising and growth, featuring speakers from Hostinger, General Catalyst, and HealthQuest Capital, will cover practical frameworks for managing rounds and leveraging agentic workflows for marketing.
Why It Matters
For developers and entrepreneurs, the breakout sessions signal a maturation of the AI conversation from general capability hype to specific implementation challenges. The inclusion of Anthropic staff detailing their internal agent workflows provides a rare, practical glimpse into how frontier model developers are integrating AI into daily engineering practices, moving beyond theoretical applications to operational reality.
The focus on the "Inference Economy" highlights a critical economic shift for the industry. As frontier models become more commoditized, the source material suggests that infrastructure providers and startups must rethink their moats. The discussion around latency as a product feature underscores the growing importance of specialized hardware and cloud infrastructure decisions for companies that do not own the underlying model. This shift impacts how investors evaluate startups, potentially favoring those with strong infrastructure or data advantages over those relying solely on model access.
Furthermore, the dedicated sessions on physical AI and trust indicate that the industry is grappling with the transition from digital-only intelligence to systems that interact with the physical world. By bringing together builders and investors from firms like Cerebras, SOSV, and The Aerospace Corporation, the agenda reflects a concerted effort to define standards and identify opportunities in sectors requiring real-time processing and autonomous decision-making.
The Bottom Line
TechCrunch Disrupt 2026’s breakout agenda prioritizes actionable insights over high-level predictions, with limited-capacity sessions designed for direct engagement. Attendees can register for the October 13-15 event to access these discussions on agent workflows, inference economics, and physical AI, with discounts available for early registration and group purchases.