TechCrunch Disrupt 2026 will host a session exploring the strategic choice between open-source and proprietary AI models for startups building AI products. This decision is fundamentally important because it affects cost, infrastructure, control, differentiation and speed—making it one of the most consequential choices founders face in the current AI landscape. The session, led by Nvidia executives Nader Khalil and Sydney Sykes, will examine the trade-offs and whether either approach offers lasting competitive advantage.

Both open and proprietary AI models have advanced significantly, making the choice less about feasibility and more about commercial sense. Nvidia reported in July that 145 papers at ICML 2026 cited its open-source Nemotron models, demonstrating strong adoption across robotics, autonomous vehicles and biomedical research, whilst proprietary frontier labs continue pushing capabilities forward. Nvidia's CEO Jensen Huang has argued that the future lies in combining both approaches rather than choosing exclusively one, a perspective reflecting the complexity startups now face when deciding how tightly to tie their products to any single model.