Open or closed AI? How founders are choosing what to build on at TechCrunch Disrupt 2026

The cost to build and deploy custom AI models falls significantly across the board.

TechCrunch Disrupt 2026 will host multiple discussions on how AI startups are navigating the evolving landscape of model choices, focusing on the decision between open-source and proprietary AI. Founders are increasingly adopting multi-model strategies, customizing existing models, and considering hardware implications. Sessions will feature experts from CapitalG, Together AI, Pathway, Oumi, Nvidia, and Riccursive Intelligence. These discussions will explore the trade-offs between cost, performance, flexibility, and ownership in AI development. The event, taking place October 13-15 in San Francisco, aims to provide founders with practical frameworks and insights for making architectural decisions that shape their products and businesses. It highlights that model choice is no longer a static decision but a dynamic one influenced by rapid technological advancements and market demands.

Founders are presented with an expanding toolkit for AI development, moving beyond a singular reliance on proprietary APIs. The emergence of robust open-source models and advanced customization techniques allows for more nuanced and potentially cost-effective solutions. This strategic flexibility enables startups to tailor AI capabilities to specific workloads, optimize performance, and adapt quickly to new advancements, impacting operational costs, product roadmaps, and competitive differentiation.

The conference will delve into the intricate relationship between AI model architecture and the underlying hardware, suggesting that future AI development will be increasingly co-designed. This convergence implies that hardware innovation may accelerate, directly influencing the speed at which new AI capabilities become accessible to startups. Understanding this synergy is crucial for founders to anticipate market shifts and position their companies for long-term success in a rapidly evolving technological ecosystem.