New: The State of Fitness, Q2 2026. Get the free report
Capital and M&A

Who Has the Right to Win? The Health Data Platform Race

The winners in health data won’t be the companies with the best dashboard. They’ll be the ones that own something their competitors can’t rent.

Key takeaways

  • Shared data feeds and rented AI models make most health dashboards easy to copy.
  • Eric sees four durable advantages: proprietary data generation, a regulated last mile, trusted distribution, and human accountability.
  • Gym operators already have distribution and accountability. Before taking a free platform deal, ask who owns the member data and what happens if funding falls through.

Eudemonia is a strange event, and I mean that as a compliment. In one hallway you have people who are very clearly on mushrooms. In the next you have founders and investors doing serious business. Tom Brady on one stage, Andrew Huberman on another.

I spent three days there, and no matter who I sat down with, one category kept coming up: the all-in-one health data platform. The pitch is the same every time. Wearables in, labs in, medical records in, multiomics if you want to get fancy. An AI layer on top. Sometimes a practitioner in the loop, sometimes not. Different logos, same platform.

Between the summit and my own calendar, I have now taken meetings with at least a dozen founders building in this category. I ask all of them the same two questions. Who has the right to win here? And what stops someone from skipping your platform entirely and asking Claude?

Here's where I landed.

The Stack Is a Commodity

Start with what these platforms are made of, because almost none of it is proprietary. The wearable data arrives through the same handful of SDK rails. The AI is rented from the same three model companies. The dashboards are nearly identical - rings, scores, and a chat window. Even the blood work mostly runs through the same places. Function Health, the biggest name in consumer testing, draws its samples at Quest Diagnostics locations.

There are exceptions on the testing side. Vero Bioscience built its own organ-aging assay. A few companies run proprietary panels you cannot order anywhere else. Hold that thought, because owning the test turns out to be one of the few real moats in this story.

JJ Creegan said something on my show a few months back that applies to the whole category. Brands are taking their customer data and loading it into Claude, and "you're paying for the same hundred dollar per month membership to Claude that I am."

When every layer of the stack can be rented, the stack is not a strategy.

Different Doors, Same Building

Founders in this category love the word wedge. The wedge is the product that gets you in the door, the thing you sell today so you can sell everything later.

The testing company told me the test is a wedge. The records company told me the records are a wedge. Ask where the wedge leads and every founder describes the same destination: an end-to-end platform running from your data to your diagnosis to your intervention. Nobody is building a blood test or a health app. Everybody is building the same future company and selling a different door into it.

The most experienced platform operators I met, people with decades of enterprise channel work behind them, were the only ones who refused the full-stack ambition. No medical advice, no regulated territory, stay a tool the practitioner controls. One of them summed up the category's problem in two sentences: the best thing about the product is the unlimited use cases, and the worst thing about the product is the unlimited use cases.

The veterans treat focus as the scarce asset. The newcomers treat it as a constraint to escape.

The Claude Question

230 million people ask ChatGPT health and wellness questions every week. I quote this number constantly from their January announcement.

So I ask founders directly: what happens when the consumer brings all their data to a general AI assistant instead of your platform?

The honest ones do not claim they will out-interpret it. They say the assistant wins the conversation, and their job is to be the thing it plugs into. There is plumbing behind that worth knowing about: MCP, the Model Context Protocol. It is the standard that lets an AI assistant reach into an outside platform, pull your context, and take action on your behalf. The sharpest founders are already building MCP-ready platforms so that whatever assistant you use can call theirs.

I think they are reading it right, and the implication is hard. The interpretation layer - the scores, the insights, the explanations of your biomarkers - is heading to zero. What stays valuable is what sits on either side of the insight.

The Four Moats

Upstream of the insight: data generation. If creating the data requires atoms - a sensor on a wrist, a proprietary assay in a lab, a blood draw in a chair - a model company cannot commoditize it. This is why the testing exceptions matter. Owning the assay is a moat. Reselling someone else's is not.

Downstream of the insight: the regulated last mile. A general assistant cannot order a panel, prescribe, carry malpractice liability, or bill insurance (yet). Whoever owns the licensed step between a flagged number and a treated person owns something durable.

Beside the insight: distribution and trust. The practitioner, the employer, the gym floor. Whoever the person already believes when they hear "you should get this checked."

And underneath everything: accountability. A human who notices you did not show up this week and says something. No platform has replaced that, and if you read this newsletter, you get it.

A platform that owns one of these four has a right to win. A platform that owns none of them - SDK data plus rented AI plus a dashboard - is the commodity middle. I have spent three years watching the middle get squeezed in every corner of this industry. This category will not be the exception.

The Practitioner Question

The no-practitioner platforms are pure software. Cheaper, faster, app-store distribution. They are also in a head-on collision with the assistant, because a chat window explaining your data is exactly what Claude already does, with more of your context and no new subscription.

The practitioner-in-the-loop platforms are slower. Selling into practices takes quarters, not days. But look at what they collect along the way: trust, the regulated last mile, and distribution. Three of the four moats in one move. And the economics can work without venture math, because for a cash-pay practice, the platform can cost less per patient than a single office visit brings in. The practice pays for itself, and the patient has a reason to stay.

My read: the durable positions in this category will have a licensed human somewhere in the loop. The pure-software middle is where the body count will stack up (figuratively...I hope).

Who Is This For?

Ask who these platforms are built for and you get a list: consumers, practitioners, employers, gyms, and, said out loud more often than you would expect, investors. When the audience list has five entries, the business model has not been chosen yet.

Corporate wellness shows up in almost every deck and survives almost none of the scrutiny. The sales cycles run through brokers and benefits planners and take ages, as one founder put it. Jon Canarick told me on the show what years of pitching payers taught him: most employees do not stay on one insurance plan long enough for prevention to pay the payer back.

The more interesting answer is specific populations. Two different companies told me, unprompted, that their product lands hardest with people managing something. Menopause. Autoimmune conditions. Cancer survivorship. Chronic illness. That tracks. The general wellness optimizer on the pitch deck cover is the most crowded, least loyal customer in health tech. The person with a real condition shows up, logs the data, and comes back, because the data is attached to something that matters. These are also the people traditional healthcare serves worst between appointments.

If I were building in this category, I would pick a population before I picked a feature.

What This Means for Operators

If you run gyms or studios, you are not a bystander in this race. You are the prize. There are platforms right now offering their product to fitness operators for free, because your members and their data are worth more than your money. That is the category telling you which moats you already own: distribution and accountability.

So when one of these platforms pitches you, and they will, ask two questions. Which of the four moats do they own? And what is their capital model? A land-grab company giving you free product is buying users to justify its next funding round. That can still be a good deal, but go in knowing you are the acquisition channel, and ask what happens to your members' data if the round does not come. A platform with self-sustaining economics will still exist in three years. Those are two different partners.

A Final Note

I do not know who wins this category. After months and months of asking questions, I am not convinced the founders do either, and my guess is there will not be one winner anyway. There will be several, roughly one per moat, and a long list of identical platforms in the middle that quietly disappear.

The test fits on an index card. Do they own something that cannot be rented? The data, the regulation, the distribution, or the relationship. Everything else in the stack is available to everyone, including the person holding the phone.

Frequently asked questions

What gives a health data platform an advantage?

Eric looks for four things a competitor cannot easily rent: proprietary ways to generate data, a licensed step that turns an insight into care, distribution through people customers already trust, and human accountability. A platform does not need to own all four, but a dashboard alone is not enough.

How could general-purpose AI change the health platform market?

Eric expects general assistants to compete directly with apps that primarily explain data. A platform can still matter if the assistant can connect to it and use something it cannot provide on its own, such as a proprietary test or access to a licensed practitioner.

What should a gym ask before signing up for a free health platform?

Ask what the platform owns that competitors cannot easily copy, how it pays for itself, and what happens to members’ data if a funding round falls through. Eric’s point is that the gym already brings a relationship with its members, even if the platform is free.

More in Capital and M&A ↗

From Eric, every week

The conversations worth your time, in one email.