Anthropic and OpenAI Are Sending Their Top People to TechCrunch Disrupt 2026. Here's Why It Matters.

Key points

• TechCrunch Disrupt 2026 runs October 13 to 15 at Moscone Center in San Francisco, with a dedicated AI Stage presented by Google for Startups.
• Anthropic's Head of Applied AI and OpenAI's Head of Productivity are both confirmed speakers, focusing on real-world deployment patterns rather than demos.
• A new job title called "GTM engineer" (a specialist who uses AI to redesign how companies find and win customers) did not exist two years ago and is now one of the fastest-growing roles in tech.
• Enterprise AI security is a central theme, with speakers from AWS, Okta, and Databricks addressing gaps that current security tools were not built to handle.
• Early-bird tickets save attendees up to $200, with that pricing window closing soon.

Picture the world's most closely watched AI lab sending its top deployment expert onto a stage to answer the question nobody dares ask in a press release: why do so many enterprise AI projects stall after the demo?

That is exactly what TechCrunch Disrupt 2026 is promising. Cat de Jong, Head of Applied AI at Anthropic (the company behind the Claude family of AI assistants), will speak about what she sees once companies actually switch Claude on inside their real workflows. Not the pitch. The messy bit after.

Alongside her, Tara Seshan, Head of Productivity at OpenAI (the maker of ChatGPT), will cover a shift that has quietly reshaped hiring across Silicon Valley.

What is a GTM engineer, and why does it matter?

Two years ago the job did not exist. A "GTM engineer" (short for go-to-market engineer, meaning someone who uses AI tools to handle the whole process of finding customers, reaching them, and closing deals) is now one of the fastest-growing roles in tech. Solo practitioners are reportedly building million-dollar businesses around it.

Seshan's session will explain how AI collapsed a set of traditional sales and marketing jobs into one new discipline, and what that means for anyone building or working inside a company that sells software.

Should founders worry about AI security right now?

Yes, and several sessions exist specifically because current security tools are not keeping pace. AI "agents", meaning software that can make decisions and take actions on its own without a human approving each step, are now running inside sensitive corporate systems. The speed at which they operate is far faster than traditional security checks were designed to handle.

Arsalan Tavakoli, co-founder of Databricks (a data and AI platform used by thousands of companies), and Ric Smith, President of Product and Technology at Okta (an identity and access management company, essentially the gatekeeper that decides who gets into which system), will both address this. Smith's session focuses specifically on why letting an AI agent self-manage its own permissions is a structural mistake, not just a configuration error.

AWS, Amazon's cloud computing division, and security specialists Luta Security and 1Password also have a joint session on what enterprise-grade AI security actually requires in 2026.

What about the SaaS business model?

SaaS, short for software-as-a-subscription, is the pricing model that built much of modern tech. Charge a monthly fee, scale the user count, repeat. AI is breaking that formula because the underlying AI models are becoming cheaper and more interchangeable every quarter. If your product's secret ingredient is now a commodity, how do you justify the price?

Founders from Glean (a workplace search tool), Monte Carlo (a data reliability platform), and investors from Sapphire Ventures and NEA will tackle that live on stage.

The event also covers visual AI, with founders from Decart and Luma AI discussing what happens when AI moves beyond generating images into genuinely reasoning about what it sees in real time.

Tickets are on sale now, with up to $200 off during the current pricing window.