At TechCrunch Disrupt 2026, AI founders are navigating the complexities of choosing between open and proprietary models. The event, taking place October 13-15 in San Francisco, will feature over 10,000 founders, investors, and tech leaders.
The conference will host sessions exploring how companies can leverage multiple models for different tasks, balancing cost, performance, and flexibility. The session 'The Real Tokenmaxxing' will include insights from Mo Jomaa, Vipul Ved Prakash, and Zuzanna Stamirowska.
Founders are also grappling with how much of the AI stack they should own, with Manos Koukoumidis offering a practical framework for decision-making. The session 'Which AI Should Your Company Actually Deploy' will provide attendees with three decision principles for their next architecture conversations.
Nader Khalil and Sydney Sykes will examine the trade-offs between frontier APIs and open-weight models, highlighting their impact on product strategy and long-term differentiation. The session 'Building AI Startups Worth Betting On' will explore these choices from both builder and venture perspectives.
The event will also address the growing connection between AI and hardware, with Anna Goldie and Azalia Mirhoseini discussing how AI is optimizing chip design and influencing infrastructure. These discussions underscore the importance of preserving flexibility as AI technology evolves.
Source: techcrunch
