The Gap Was Always the Data: How Tech Is Filling the Holes in Women's Health Research
In August, one of the most powerful women in tech gave her first interview after stepping away from one of the biggest jobs in the industry. Fidji Simo had been running a major part of OpenAI. She didn't leave for a bigger title. She left because her body wouldn't let her keep going.
For seven years, Simo has lived with POTS, or postural orthostatic tachycardia syndrome. It's the condition where your heart races when you stand up, your energy disappears, and your thinking goes foggy. She told Fortune her health was the worst it had ever been.
Where she's putting what energy she has is a company called ChronicleBio. Its entire premise comes down to one idea every woman reading this should sit with for a second: conditions like hers don't have cures because nobody bothered to collect the data.
That idea is the real story of women's health right now. And the most interesting people working on it are, increasingly, women building that data themselves.
The gap was never only about money
You've probably heard that women's health is underfunded. It is. Women spend nearly 25% more of their lives in poor health than men, which adds up to roughly nine years. Yet less than 5% of global health research funding goes to conditions specific to women outside of cancer.
Private money tells the same story. Women's health captures only about 6% of private healthcare investment, and around 90% of that goes to three areas: women's cancers, fertility, and pregnancy. Important, yes. But it means conditions like autoimmune disease, heart disease in women, chronic fatigue, and POTS get almost nothing.
Underneath the funding gap sits a quieter one. When researchers in the UK looked at clinical trials, they found trials enrolling only men outnumber trials enrolling only women by 67%. And when a team at the University of Colorado dug into NIH's massive All of Us research program, they found that something as basic as when a woman went through menopause is often missing from the health records researchers rely on.
Sit with that one. Menopause happens to half the population, and it shapes heart, bone, and brain health for decades. It frequently isn't written down. You can't find a pattern in information nobody recorded.
Women are building the dataset themselves
This is where ChronicleBio gets interesting. The company, which Simo cofounded in 2025, has gathered more than 3,500 tubes of blood from people with POTS, ME/CFS, and long COVID, three conditions that overlap constantly and skew heavily female. This summer it started sending mobile blood draw trucks directly to patients' homes. The first 250 people were free, and after that the test costs $400, charged at cost. Participants get a research report back and repeat testing over time.
Why trucks? Because the sickest patients are often bedridden, which is precisely why they never show up in medical studies. If you can't get to the clinic, you don't get counted.
The company pairs blood samples with wearable data and medical records, then uses AI to look for hidden subtypes of these diseases and possible drug targets. Simo's argument is simple: AI got good at language because the internet handed it an enormous library of words. Biology has no equivalent library, especially for women's bodies, so someone has to build it. A fair note of caution here: ChronicleBio is early, it hasn't published results, and $15 million is small money next to pharma budgets. But the model matters.
She isn't alone. Earlier this month, Evvy raised $40 million on the strength of what it calls the largest clinical grade dataset on the vaginal microbiome and women's health outcomes, built from more than 100,000 patients. If you've ever been sent home from an appointment with a guess about a recurring infection, this is the kind of data that could finally replace the guess.
Even period blood is getting a second look. Long treated as waste, it contains immune cells and tissue from the uterus itself. A 2026 study in the BMJ found that period blood collected on a simple mini pad worked about as well as a clinician collected sample for HPV testing, and several companies are now collecting menstrual fluid to hunt for signatures of endometriosis.
The device on your wrist is now a research tool
If you wear a smartwatch or ring, you're sitting on data researchers have wanted for decades. Harvard researchers working with the Apple Women's Health Study analyzed more than 94,000 nights of Apple Watch sleep data from 338 women around their final period. In the 18 months before menopause, 60% of them spent more time awake after falling asleep. Just as telling, some women's sleep barely changed at all. Averages have always hidden women, and continuous data is finally showing the range.
Apple added perimenopause and menopause tracking to its Health app this year, and Oura says it holds sleep data from roughly 850,000 women between 40 and 60.
Here's the smart friend caveat. A review published in August looked at 80 menopause apps and found that 75 of them offered no disclosed evidence behind their insights, and none of the major wearable brands has a published trial validating its data for perimenopause specifically. So treat your tracker as a conversation starter with your doctor, not a diagnosis. It's genuinely useful for showing patterns over months that a 15 minute appointment will never catch.
AI is shortening the wait
The biggest promise of all this data is speed. Endometriosis affects more than 190 million women worldwide and typically takes seven to ten years to diagnose, often requiring surgery to confirm. This month, researchers at Adelaide University unveiled an AI tool called EndoFusion that reads MRI and ultrasound scans together and spots signs of advanced endometriosis in 18 milliseconds. It was right 83% of the time, better than any competing model.
It's happening across the board. Clairity Breast, the first FDA authorized AI tool of its kind, predicts a woman's five year breast cancer risk from a standard mammogram, which matters because older risk models were built largely on white European women. An Austrian startup called Aitiologic just raised funding for an AI blood test that aims to flag preeclampsia risk early in pregnancy. And AI models for PCOS keep posting impressive accuracy, though most studies are still small, so the results look stronger on paper than they've proven in real clinics.
Investors have noticed. Women's health companies using AI are now valued at nearly three times the sector median, and that money is flowing mostly into tools that predict and prevent serious health events.
The institutions are finally moving
The public side is catching up, too. In February, the FDA removed cardiovascular, breast cancer, and dementia warnings from the boxed labels on six menopausal hormone therapy products. On September 14, NIH announced $21 million to build computer models of how hormones regulate themselves across a woman's life, so future drugs can account for female biology from the start.
Three days later, the FDA held its first public workshop in about two decades on testosterone for menopausal women. There is still no FDA approved testosterone product for women, even though many use it off label. The agency received hundreds of comments, many from women describing real changes in energy and quality of life. If you have a view, the public comment period is open through October 19 on Regulations.gov. Search "testosterone use in menopausal women."
Philanthropy is pushing hard as well. Wellcome Leap and Melinda French Gates's Pivotal have now committed more than $250 million to women's health research, including programs on heart disease in women, women's Alzheimer's risk, and heavy menstrual bleeding.
The catch: AI can repeat the old dismissal
Now for the part that should keep all of us paying attention. Researchers at MIT found that popular AI models, including GPT 4, recommended lower levels of care for female patients and were more likely to suggest they manage symptoms at home. A London School of Economics study found some models wrote softer descriptions of women's physical and mental health needs than they did for identical cases with the gender swapped.
That's the whole point in miniature. AI trained on decades of medicine that underweighted women will underweight women, only faster. Which is why the real win in this moment isn't the algorithm. It's the data. Women contributing their blood, their cycles, their sleep, and their symptoms are rewriting the record the algorithms learn from.
Back to POTS and what to do if this sounds familiar
POTS affects an estimated 1 to 3 million Americans, and around 80% of them are women. The symptoms read like a list of things women get told are stress: exhaustion, brain fog, dizziness, a pounding heart. Many women hear "anxiety" for years before anyone checks.
Here's what the research actually shows. Women with POTS seek care as often as men and report anxiety at similar rates, which means the long diagnostic delays point back to the exam room, not the patient. There's still no FDA approved drug for POTS. And there are no studies on what happens to POTS as women move through perimenopause and menopause, even though hormones clearly influence symptoms.
If this sounds like you, a few practical steps help. Use your wearable to note your heart rate lying down and then after standing for several minutes; a sustained jump of 30 beats per minute or more is one of the core signs clinicians look for. Bring that record to your doctor and ask about an active stand test or a tilt table test. Dysautonomia International is a solid place to find specialists and patient education.
When Simo talks about why she's doing this, her point is that if AI can do almost anything but can't help cure disease, it will have missed what matters most. For generations, women's bodies have been the missing chapter in medicine. The difference now is that women are the ones writing it, one blood draw, one night of sleep data, one logged symptom at a time. Your body is your biggest investment. It's time the research treated it that way.
EGL’s Take
The women's health gap was never only about money. It was about data. Medicine never collected enough information about women's bodies to see the patterns, so conditions like POTS got filed under anxiety. What's changing is who is building the dataset. This is the female longevity correction in its most concrete form: women turning their own biology into the evidence medicine skipped.
