Anthropic and OpenAI Headline TechCrunch Disrupt AI Stage in October

TechCrunch Disrupt 2026 will run October 13 to 15 at Moscone Center in San Francisco, and the event's AI Stage now has two headline speakers confirmed: Cat de Jong, Head of Applied AI at Anthropic, and Tara Seshan, Head of Productivity at OpenAI. The stage, presented by Google for Startups, focuses on the practical problems founders are hitting after deployment, including how to price AI products when models commoditize, how to secure agentic systems, and what "AI-native go-to-market" actually looks like in practice.

What happened

Detail	Fact
Event	TechCrunch Disrupt 2026
Dates	October 13-15, 2026
Location	Moscone Center, San Francisco
Stage sponsor	Google for Startups
Expected attendance	10,000+ startup, tech, and VC leaders
Ticket savings available	Up to $200 (current pricing window closing)

TechCrunch confirmed two anchor speakers for the AI Stage. Cat de Jong (Anthropic’s Head of Applied AI) will speak about what enterprise Claude deployments actually look like after the pilot phase: where they succeed quickly, where they stall, and what separates teams extracting real value from those still running trials 18 months in. Tara Seshan (OpenAI’s Head of Productivity) will cover how AI has reshaped the go-to-market function and created the new “GTM engineering” role, which according to TechCrunch did not exist two years ago and now supports independent practitioners building million-dollar businesses.

Several other sessions are confirmed alongside those two:

• Enterprise AI security: Arsalan Tavakoli, Co-founder and SVP of Field Engineering at Databricks, on observability, governance, and architecture for trustworthy enterprise AI.
• Agent security: Ric Smith, President of Product and Technology at Okta, on why application-level permission models are “fundamentally flawed” and what infrastructure-level agent security actually requires.
• Video AI: Dean Leitersdorf (Decart) and Amit Jain (Luma AI) on real-time inference and physical reasoning at the frontier.
• SaaS business models: Arvind Jain (Glean), Barr Moses (Monte Carlo), Cathy Gao (Sapphire Ventures), and Aaron Jacobson (NEA) on pricing AI products sustainably and building defensible moats when models commoditize.
• GTM engineering deep-dive: Kareem Amin, Co-founder and CEO of Clay, on the specifics of the role.
• Cloud security: Chet Kapoor (AWS), Katie Moussouris (Luta Security), and Wendy Nather (1Password) on the infrastructure view of enterprise AI security in 2026.

Why it matters

The speaker list reflects where enterprise AI pain actually sits right now. The “what model should we use” conversation is largely over. The live questions are about what happens after you ship: how do you govern an agent that is making autonomous decisions inside sensitive systems? How do you price a product when the underlying model cost is falling fast? How do you build a sales motion for software that did not exist as a category 18 months ago?

The inclusion of security-focused sessions from Okta, Databricks, AWS, and Luta Security is notable. Agentic AI (systems that can take actions autonomously, not just answer questions) creates attack surfaces that traditional identity and permission frameworks were not designed for. This is increasingly a blocker for enterprise deals, and it is getting dedicated stage time for good reason.

For anyone tracking how AI integration decisions play out inside real businesses, hearing directly from the people at Anthropic and OpenAI who manage those deployments is more useful than another product announcement. The Anthropic session in particular promises post-pilot data: patterns from actual production workflows, not slide-deck projections.

Our take

The topic framing is sharper than most conference marketing. “What separates teams extracting real value from ones still running pilots after 18 months” is exactly the question we hear from clients. Most teams we talk to are somewhere in that gap: they have a working prototype, but turning it into something reliable, auditable, and worth paying for is a different problem entirely.

The GTM engineering angle is also worth paying attention to. We have covered the structural pressure AI puts on headcount, and GTM is one of the areas where roles are being rebuilt rather than just cut. If your business relies on a traditional sales or marketing stack, the Seshan and Amin sessions will give you concrete language for what is replacing it.

That said, conference sessions are still conference sessions. The most useful thing to bring away from an event like this is not the keynote take but the hallway specifics: what stack are people actually running, what vendors got quietly dropped, and what did the pilot actually cost. The stage is a starting point.

What to do about it

1. If you are attending Disrupt, prioritise the de Jong (Anthropic) and Smith (Okta) sessions if enterprise deployment or agent security are live concerns for your team.
2. Check the current ticket pricing before the discount window closes; TechCrunch states savings of up to $200 are still available.
3. If you cannot attend, watch for post-event session recordings. TechCrunch typically publishes them.
4. If the SaaS pricing question is live for your business, the Glean, Monte Carlo, and Sapphire panel is worth your time. Pricing AI products when model costs are falling is one of the harder strategic problems right now, and you will want frameworks from people who have shipped real products.

If you are working through your own AI deployment questions before October, the Lumien team is available to talk through what the post-pilot stage actually looks like.

Source: TechCrunch · AI