At a recent StartUp Health Fireside Chat, Unity Stoakes sat down with Kimmi Grewal, Vice President of AI Ecosystem Partnerships at Microsoft, to talk about a shift she sees playing out across the health innovation ecosystem: the model is no longer where the advantage lives.
Watch or listen to an excerpt
Grewal, who leads Microsoft's partnerships with frontier AI labs and ecosystem providers – and who has served on the StartUp Health Impact Board since 2021 – has a front-row seat to how founders are turning AI's potential into real products. Her read on the current moment is direct. Venture investment in AI during the first quarter of this year alone surpassed the full-year total for 2025, and open models are closing the performance and cost gap with frontier labs fast enough that more startups and enterprises are building on them by default.
That shift changes where the real differentiation sits. As Grewal frames it, the leaderboard for top models changes too often for any startup to build a durable edge on model choice alone. The winners are the ones owning a specific workflow and a proprietary data set – and, just as important, guarding that data so it never becomes training material for someone else's model.
One of the more counterintuitive points Grewal raised: healthcare isn't lagging behind in AI adoption the way regulated industries typically do. It's moving as a fast follower, in some cases ahead of sectors expected to move first. Long-standing, well-understood problems in healthcare finally have a real shot at solutions, which is driving faster experimentation even as production deployment still requires the extra step of safety and governance review.
Are you ready to be in the room?
Become a StartUp Health member to catch fireside chats like this one live, ask your own questions, and connect directly with the investors and health leaders shaping what's next.
For founders, Grewal's advice comes down to sequencing. Start with the customer and the problem. Only then decide which tools – AI included – actually solve it. Build a partnership strategy early rather than treating it as an afterthought. And design architecture flexible enough to swap models without losing your edge, because betting everything on today's top performer is a bigger risk than it looks like.
She also pointed founders toward Microsoft for Startups as an accessible front door – offering up to $150K in Azure credits – with Pegasus as a more selective, hands-on track for startups showing strong enterprise alignment and marketplace traction.
Grewal closed with a note aimed at anyone who feels like they're behind: the tools available today make it possible to teach yourself what you need to compete, no matter when you started. And for founders building something focused and real, she sees the path to an eventual exit as more open than it's been in years, as enterprises look to prove out their own AI investments by partnering with startups who've already done the hard part.
🚀 This post covers a portion of a live StartUp Health Fireside Chat. StartUp Health members get access to full conversations and the opportunity to ask questions in real time. Join StartUp Health to connect with Health Transformers, investors, and partners, and sit in on fireside chats, expert AMAs, and more.
