Disrupt 2026 to Probe AI Founders' Open vs Closed Choices — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- TechCrunch Disrupt 2026 (October 13–15, Moscone West, San Francisco) will run four panels on how AI founders choose between open and proprietary models, with more than 10,000 attendees, 250+ speakers, and 300+ exhibiting startups expected.
- CapitalG's Mo Jomaa, Together AI's Vipul Ved Prakash, and Pathway's Zuzanna Stamirowska will debate multi-model strategies on the "Real Tokenmaxxing" panel, examining how companies balance cost, performance, and flexibility across different models for different jobs.
- Oumi CEO Manos Koukoumidis will lead "Rent, Customize, or Build," guiding attendees through frontier APIs, customized open weights, and owning AI outright, using audience polls and a framework yielding three decision principles for architecture conversations.
- Nvidia's Nader Khalil and Sydney Sykes will examine what founders are choosing today between frontier APIs and open-weight models in "Building AI Startups Worth Betting On," framing the choice as one that shapes product strategy and long-term differentiation.
- Ricursive Intelligence founders Anna Goldie and Azalia Mirhoseini will argue in "When AI Starts Designing Its Own Hardware" that AI is already optimizing chips, tightening the connection between model architecture and the silicon underneath it.
Why it matters: The build-vs-rent question for AI is no longer just an API contract — the source maps it from model selection down to chip design, with explicit trade-offs in operating costs, product control, and the ability to create differentiation competitors can't replicate. Founders attending will leave with decision frameworks for whether to rent, customize open weights, or own the stack outright.
Ask SkimNews
