Anthropic and OpenAI Headline the AI Stage at TechCrunch Disrupt 2026
Anthropic and OpenAI will appear on the AI Stage at TechCrunch Disrupt 2026, joining a speaker lineup that covers enterprise deployment, agent security, and the practical realities of building AI-native organizations. The conference runs October 13–15 at Moscone Center in San Francisco (TechCrunch, 2026-08-27).
Cat de Jong, Head of Applied AI at Anthropic, will present a session titled "What Anthropic Sees When Enterprises Actually Deploy Claude." The title suggests a focus on real-world usage patterns rather than product roadmap announcements. Tara Seshan, Head of Productivity at OpenAI, will present "What Building AI Native Actually Means," a framing that pushes past the now-commonplace "AI-first" label toward something more structural — how an organization is built around AI from the ground up, not just enthusiastic about adopting it.
The AI Stage is presented by Google for Startups. It is one of six industry-focused stages at Disrupt 2026, and three of those six are dedicated to AI topics (TechCrunch, 2026-08-26). The broader agenda includes speakers from Google, Amazon, and Replit alongside OpenAI, Anthropic, and Databricks.
Two additional sessions round out the enterprise-facing programming. Arsalan Tavakoli, Co-founder and SVP of Field Engineering at Databricks, will present "The Enterprise Isn't Broken. Your Assumptions About It Are." Ric Smith, President of Product & Technology at Okta, will present "The Agent Security Problem Nobody Is Talking About." Smith's session is notable for placing identity and access management — the systems that control who can do what inside a company's digital environment — at the center of the conversation about autonomous AI agents, a convergence that has been accelerating as these agents begin operating inside corporate networks.
Beyond the AI Stage, Disrupt 2026 includes a Builders Stage focused on practical strategies for scaling startups (TechCrunch, 2026-07-01). The event also features a Real World AI Stage with sessions on robots, automated factories, and extinct animals (TechCrunch, 2026-08-05). That stage sits at the more tangible end of the AI spectrum, covering physical and industrial applications rather than software-only systems.
The conference is expected to bring together 10,000 founders, VCs, and operators.
The session titles themselves tell a story worth pausing on. De Jong's "when enterprises actually deploy" framing implies a gap between what vendors show in demos and what happens when systems go live. Tavakoli's "your assumptions about it are" challenges the tendency in the AI community to treat existing enterprise data infrastructure as an obstacle rather than the foundation it actually is. Seshan's "what building AI native actually means" suggests the term has been used so loosely it now needs redefinition. And Smith's "agent security problem nobody is talking about" points at a genuine blind spot: the identity, authorization, and audit surface area that autonomous agents create inside enterprise systems — which extends well beyond the prompt-injection and data-leakage concerns that dominate current security discussions.
For practitioners attending or tracking the event, the programming leans toward operational maturity rather than model announcements. The Anthropic and OpenAI sessions are positioned around deployment experience and organizational transformation, not capability benchmarks. That is a shift from last year's conference circuit, where frontier model releases and capability demonstrations dominated. The presence of Databricks and Okta alongside the two model providers reinforces an enterprise-first framing: data infrastructure and identity management as the enabling layers for AI adoption, not afterthoughts to it.
The broader context here is one of normalization. Disrupt 2026's decision to dedicate half its stages to AI reflects the degree to which the technology has moved from a specialized track to a horizontal concern across product, security, infrastructure, and operations. For the expected 10,000 attendees, the relevant question is less whether AI will matter to their work and more how to navigate the gap between demos and deployed systems, between "AI-first" as a slogan and as an organizational reality.
