Open or closed AI? Nvidia’s Nader Khalil and Sydney Sykes take on one of the decisions shaping next-gen startups at TechCrunch Disrupt 2026
TechCrunch TechCrunch Events ● Covered by 6 sources
Nvidia’s Nader Khalil and Sydney Sykes will tackle open vs. closed AI at Disrupt 2026. Startups are choosing between speed, control, and cost, and the wrong call can shape the whole company.
Based on reporting by TechCrunch, TechCrunch Events — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Open or closed? For founders building AI products, that isn’t an abstract debate. It’s a stack-level decision that can affect cost, margins, infrastructure, speed, data control, and whether a product can actually stand apart from the next competitor using the same model.
That is the point of “The Open vs. Closed AI Debate Is Just Getting Started,” a Builders Stage session at TechCrunch Disrupt 2026 in San Francisco on October 13-15. Nvidia’s Nader Khalil, its director of developer tech, and Sydney Sykes, the company’s global head of VC partnerships, will talk through the trade-offs and the question that keeps coming back: can either open or proprietary AI create lasting advantage?
The market has already made the issue messier. Nvidia said in July that 145 papers accepted at ICML 2026 cited its Nemotron open models and datasets, and that research using its open model families was showing up in robotics, autonomous vehicles, and biomedical work. At the same time, proprietary frontier labs keep pushing capability forward. The simple either-or argument is fading, and even Nvidia’s own message from GTC earlier this year was that the future is proprietary and open, not one or the other.
But that neat slogan gets complicated once a company has to ship something. If two models produce similar results, does cheaper win? If one offers more control over data, does that outweigh the work of running your own infrastructure? And if the best model keeps changing every few months, how hard should a startup lock itself to any one approach? Khalil brings the builder side of that problem. Before joining Nvidia, he co-founded Brev.dev, which Nvidia acquired in July 2024. Brev.dev focused on making GPU infrastructure easier to use across different environments, and Nvidia’s docs say its tools can deploy AI software across public cloud, private cloud, and on-premises setups without tying developers to one compute source.
Sykes adds the investor angle. Together, the two are set up to show how technical choices turn into business choices, especially when a model is not the moat, or at least not the whole moat. Nvidia is also pushing open models itself, including Nemotron 3 Super, a 120-billion-parameter model launched in March for agentic workloads. The more interesting part may be that many companies are already mixing open and proprietary models instead of pledging loyalty to one camp.
My take — AI-written commentary, not fact-checked reporting
The open-vs-closed fight has become a little too holy for something this practical. Most startups do not need a model religion; they need lower bills, less lock-in, and a product customers will pay for. The smart play is usually the messy one: use whatever works, and keep the escape hatch open because the model market changes faster than founders like to admit.
Read more about this at: TechCrunch