TechCrunch Disrupt 2026 spotlights open vs closed AI choices for startups

In short: More startup founders are mixing and matching AI models, and TechCrunch Disrupt 2026 will focus on how to choose between open and closed AI.

What's going on

Startups building with AI are no longer treating the choice of an AI model as a one-time decision. Many teams now use more than one model, picking different tools for different tasks, like using different appliances for cooking.

TechCrunch says this comes from two changes happening at the same time. Open models are getting better, meaning the underlying model can be used and changed by others. At the same time, paid “frontier” AI services are also improving. These are AI models you access through an API, which is like ordering power from a utility instead of building your own generator.

At TechCrunch Disrupt 2026, several sessions will cover these trade-offs. One panel will discuss “multi-model” products, with speakers including Mo Jomaa of CapitalG, Vipul Ved Prakash of Together AI, and Zuzanna Stamirowska of Pathway. Another talk from Manos Koukoumidis of Oumi will focus on whether companies should rent AI, customize an existing model, or build their own.

The event will also include a session on AI and hardware, with Anna Goldie of Ricursive Intelligence and Azalia Mirhoseini discussing how AI can help design chips, which are the physical parts computers use to run AI.

What to watch

For everyday users, these choices can affect cost, speed, and reliability in AI products you use at work or at home. If more companies build systems that can switch models easily, it may become simpler to improve features quickly or keep prices down when one provider gets too expensive.

Source: TechCrunch AI