Vinod Khosla on Health Moonshots in the Age of Superintelligence: Build Trust, Then Let AI Scale Care

Fifteen years ago, the StartUp Health gatherings around JPM began with a simple idea: bring builders together to accelerate what health can become. At Apollo House 2026, that long arc was on full display as Unity Stoakes once again welcomed entrepreneur, investor, and technologist Vinod Khosla of Khosla Ventures for a fireside chat on “Health Moonshots in the Age of Superintelligence.”

Khosla framed this moment as the next in a series of platform changes, from the microprocessor to the internet to mobile, then argued that AI is broader and more consequential than any of the previous waves. The pace, he suggested, is also unprecedented, not just in adoption, but in the rate of learning and iteration that entrepreneurs can now bring to hard problems.

A central theme was how to introduce AI into healthcare responsibly. Rather than aiming for full autonomy on day one, Khosla advocated for an “AI intern” model: treat AI systems like newly trained clinicians who do the work, while licensed professionals supervise and approve final decisions, including prescriptions. As trust grows, the “leash” can loosen.

That emphasis on trust came up again and again. In Khosla’s view, healthcare’s most important commodity is not speed or efficiency, but confidence, earned through proof, guardrails, and fit-for-purpose design. He also cautioned that generic systems used irresponsibly can fail in predictable ways, and that the work of founders is to shape general capability into reliable clinical tools.

In a live exchange, Esther Dyson joined Khosla and Stoakes on stage to push the conversation toward what care means beyond diagnosis, including empathy, communication, and the importance of supporting health before people reach crisis. Khosla agreed the human element still matters, and added that technology can sometimes create space for more honest conversations, especially when patients feel less judged and more able to ask questions.

The throughline was clear: the opportunity is extraordinary, but it will not fulfill itself automatically. It will take entrepreneurs willing to build for real-world constraints, clinicians willing to supervise new workflows, and a community willing to expand what it believes is possible.