For AI startups today, choosing between open and closed models is more than just a one-time decision. As the landscape shifts, many companies are exploring various AI models to effectively address different tasks. This approach enables founders to refine their strategies, balancing costs and performance while also considering flexibility as technology evolves.
At TechCrunch Disrupt 2026, taking place from October 13-15 at Moscone West in San Francisco, this ongoing dialogue will take center stage. An anticipated crowd of over 10,000 founders, investors, operators, and tech leaders will gather to navigate these choices.
During the session titled “The Real Tokenmaxxing: How the Best AI Companies Navigate a Multi-Model World”, experts like Mo Jomaa, partner at CapitalG; Vipul Ved Prakash, co-founder and CEO of Together AI; and Zuzanna Stamirowska, CEO and co-founder of Pathway, will discuss why relying on a single AI model may no longer be sufficient. The conversation will focus on how companies can effectively utilize multiple models to boost performance and cut costs.
Another key consideration for founders is how much of their AI stack they should own. This topic will be addressed by Manos Koukoumidis, CEO and co-founder of Oumi, during his session titled “Which AI Should Your Company Actually Deploy: Rent, Customize, or Build”. Through audience polls, startup scenarios, and a practical framework, Koukoumidis will help attendees evaluate frontier APIs, customized open weights, and owning AI outright – and will leave them with three decision principles they can apply to their next architecture conversation.
Ownership of the AI model can significantly influence both the product and the business. Building more of the AI stack can offer greater control and differentiation, but it can also require more time, talent, and resources compared to simply using existing models.
The ongoing debate between open and proprietary models remains a hot topic. Nader Khalil, Director of Developer Tech at Nvidia, along with Sydney Sykes, Global Head of VC Partnerships at Nvidia, will lead the session “Building AI Startups Worth Betting On”. They’ll explore the current choices founders are making, the trade-offs between frontier APIs and open-weight models, and how these decisions impact product strategies and long-term differentiation.
Model choice carries significant business implications, influencing everything from infrastructure requirements to the level of control founders have over their products. Startups must consider how their decisions can create competitive advantages that others may find hard to replicate.
As AI technology advances, the performance of these models increasingly relies on the supporting hardware. Founders will delve into this connection in the session titled “When AI Starts Designing Its Own Hardware”, featuring Anna Goldie, founder and CEO of Ricursive Intelligence, and Azalia Mirhoseini, founder and CTO. They’ll discuss how AI is beginning to optimize chips and hardware, and the implications for the infrastructure that supports AI startups today.
The sessions at Disrupt 2026 will highlight the critical balance between ownership and model choice for AI startups. As founders weigh the benefits of building their own AI stack against the use of existing models, the discussions will shed light on how these strategic decisions can shape their product offerings and competitive positioning in a rapidly evolving industry.
