Mary Camacho — A Different Kind of Vessel

What if there were a privacy-first federated AI model for women’s wellness?

What if we, collectively in femtech, owned it?

The AI models we are all leaning on are insufficient. Anyone who has asked one about their cycle, their hormones, their changing body knows this. The infrastructure that would do better has not been built. It could be.

The wellness layer

Companies are good at producing specific benefits. There is a different category of value that asks for a different kind of vessel.

That distinction matters most when the thing being built will shape what gets seen and what stays invisible for women around the globe.

Public health, fitness, and nutrition research has existed for a long time. Women have been under-represented in that research for almost as long. Drug trials that excluded women of reproductive age. Cardiac research conducted on male subjects and generalized as universal. Pain, exercise, nutrition, sleep, all calibrated to bodies that were not the bodies most of the world’s population lives in.

The gap is not new. The literature on it is large. The clinical and research correction matters and is happening, slowly, where it can.

However, there is another layer the clinical correction cannot reach. The data that matters most for women understanding their bodies in daily life does not necessarily live in clinical or research settings. It lives in the wellness domain. Cycle patterns observed across years. Energy and mood through hormonal phases. Sleep changes through perimenopause. The body literacy women build by living in their own bodies and tracking what they notice.

This data exists in millions of women’s apps and journals and devices. It is held by the operators who built those tools, and on occasion by the women themselves. It does not exist in any research database, and certainly not in hospital records.

The wellness layer is where vast amounts of data live. It is also where it has always been going to need to come from if AI is going to represent women’s bodies as they actually are. In transition through life. In daily rhythm. In the texture of what women themselves notice and track.

AI products are now being built on top of all of this. The big AI companies are training on whatever they can license, scrape, or partner their way to. The largest operators in wellness are beginning to train on their own data, which is a step forward although, only a slice.

Every operator’s data is shaped by who they sell to. Rings do not work for women with arthritis in their fingers. Wearables come at price points many cannot afford, in form factors many do not want, from companies many do not trust with their data. Fertility apps describe women tracking fertility. Period trackers describe women who menstruate. Every company’s data defines its users. None are the population. The combination of them begins to approximate it.

So, what if we built it together?

The shape of what could be built

A cooperative of femtech and women’s wellness operators who collectively own the AI infrastructure that serves their work. What if we built it?

Each member contributes mathematical signal from inference run on their own user data or from private user compute instances, never contributing the data itself, to a federated training process. The cooperative owns the resulting models and weights as collective IP. Members get access to the resulting models on preferential terms. Non-member operators license commercially. The protocols, reference implementations, and supporting tooling are open-sourced. The trained models are member-owned.

The cooperative is domiciled in Europe, where the regulatory ground supports it and the data is protected from political pressures elsewhere. It is governed by its members.

The cooperative is positioned in the wellness space, not the clinical space, at least to begin. This is a deliberate choice with operational consequences. Wellness data sits outside the heavy regulatory regimes that govern clinical AI. It is not Protected Health Information under US law. It is not inside the clinical scope of the European Health Data Space.

The cooperative can launch in a regime designed for it, with appropriate transparency obligations under the EU AI Act, without fighting clinical conformity assessments it does not need to fight. The clinical layer can come later, by deliberate progression, with credentialed partners. Starting in wellness is the choice that makes the structure possible to build immediately.

Building this once, together, is a different proposition than every operator building its own. The compute, the engineering hours, the energy that frontier-scale model training requires, all of it spent in parallel across companies producing proprietary models that each describe only a slice. The duplicated cost produces fragmented results.

The cooperative form is what makes one good model possible instead of many inadequate ones. The shape of the answer matches the shape of the problem.

The form is borrowed from a long lineage. The Associated Press has held news distribution as a cooperative for nearly two centuries. Visa was a cooperative of member banks for thirty years before its IPO. Land O’Lakes for dairy. Ocean Spray for cranberries. Mondragón for cooperatives at federation scale. In the digital age, the MELLODDY consortium proved federated learning works at industry scale across ten of the world’s largest pharmaceutical companies. OHDSI runs federated health collaboration across more than seventy countries.

What these structures share is member ownership of the work and the value it produces. Each is governed by the entities that contribute to it, on terms they negotiate together.

What would be new is the application. A women’s wellness operator cooperative, built privacy-first by architecture, oriented to representing this population correctly in the AI infrastructure of the coming decades.

The form is borrowed. The choice to apply it here, now, with these values, is what would be new.

The Women’s Health Visibility Alliance, formed in early 2026 by Essity, Clue, Hertility, Daye, and Mooncup to challenge censorship of women’s health content on social media, showed that operators in this sector will coalesce when there is a coordinated reason to.

The reason for a wellness AI cooperative is larger. It is about who gets to be visible in the AI products of the next decade, and on what terms.

How I came to be writing this

I started saying around 2019 that the place I most wanted to see distributed systems work applied was in health data. The intuition was specific even when the articulation was not.

Since then, I have been working out what that means.

At Holochain, the work was on the infrastructure layer, on distributed and peer-to-peer systems. The conversations with colleagues were about how privacy-preserving architectures could carry information that mattered without aggregating power that shouldn’t be aggregated. Healthcare came up repeatedly because the values fit was so clean.

Through my work, the frame has been clarified.

That sentence is the destination of seven years of thinking. It was not the starting point.

I am not in clinical health tech and never have been. The questions I wanted to work on were closer to daily life than to the doctor’s office. That is how I came to the consumer wellness layer, and to building Cirdia.

Cirdia is a privacy-architected intelligence engine for women’s wellness. The apps we are building run pattern intelligence without centralizing user data. The architecture is privacy-first because years working at the infrastructure layer made it obvious that it was the only way for a company to keep a promise.

What I am doing

This is not an announcement of a coalition. It is not a brand. It is the start of conversations, in earnest, with people who might find this compelling enough to do the work of helping it emerge.

The work is real. Convening founding members. Designing governance that holds across borders and over decades. Writing privacy and consent standards. Spinning up the technical foundation. Securing anchor funding from grants and philanthropic capital willing to underwrite mission-aligned infrastructure.

None of this is fast. None of it is easy. The cooperatives that have lasted have all gone through years of careful work before their value became visible.

The people I am gathering are people who would do that work.

Operators in women’s wellness whose companies could be founding members.

Technical leaders in federated learning, privacy-preserving computation, and cooperative infrastructure who have built this kind of system before.

Philanthropic capital and program officers whose work is anchoring public-interest infrastructure that markets cannot build alone.

Convening authority in adjacent communities, in trade alliances, in women’s health policy, in AI governance, that touches this work.

If you are one of those people, I would value being in conversation with you. The brief I am working from is a document I share with people who want to see it. Not as a finished plan. As what I have so far, open to the people who would help shape it into something better.

You can reach me on LinkedIn or send me a message here.

The infrastructure that will represent women’s wellness in the AI of the next decade will be built one way or another. The question is by whom, and on what terms.

What I am doing is opening the conversation. What happens next depends on who joins it.

privacy-architecture ai femtech cooperative