Delfina wants pregnancy care to predict, not react.

The United States spends more per person on healthcare than any country on earth, and it has the highest maternal mortality rate in the wealthy world - more than double the rate of comparable high-income nations. For Black women the gap is starker still, with maternal mortality roughly three times that of white women. This is the kind of statistic that people quote solemnly and then move past, because it is hard to know what a normal person, or a normal company, is supposed to do about it.

Delfina's answer is that a lot of the harm happens in a place nobody is looking: the weeks between prenatal visits. A patient sees her OB, everything looks fine, and then she goes home for three or four weeks during which her blood pressure or her blood sugar or her weight does something the doctor would very much have wanted to know about in real time. Traditional prenatal care is built around discrete appointments. Delfina is built around the gaps.

The company, founded in San Francisco in 2021, sells what it calls the Delfina Care Platform. On the clinical side it is software: proprietary predictive models that ingest a patient's data, compute a risk profile, and tell the care team which patients are drifting toward preterm birth, a hypertensive disorder, or gestational diabetes - the conditions that drive most maternal and infant harm. On the patient side it is an app - a pregnancy tracker where you log symptoms, mood, weight, blood pressure, and blood sugar, sometimes through connected cuffs, glucometers, and scales that the provider sends home with you.

The clever part, and the part that separates Delfina from the enormous pile of pregnancy apps that count kicks and sell onesies, is that the two halves are wired together. Your data does not sit on your phone forming a nice chart. It goes to a model, and if the model gets nervous it goes to your actual clinic, and a human on Delfina's care team - a doula, a dietitian, a mental health therapist, a lactation consultant - can reach you. The doulas are bilingual, English and Spanish, and available around the clock. It is, in a sense, the least glamorous idea in health tech: make the information reach the person who can act on it, at the time they can still act.

The founder had seen the failure mode up close

Delfina's CEO is Dr. Senan Ebrahim, who holds both an MD and a PhD in computational neuroscience from Harvard. That is a fairly specific combination - it is roughly the résumé of someone who could build the risk model and also read the ultrasound. He has said he co-founded the company after experiencing, up close, the challenges that expecting parents and their providers face, including a preventable stillbirth witnessed during his training. Before Delfina he founded Hikma Health, a nonprofit building digital health tools for refugees and migrants, which tells you something about where his instincts point: toward the patients the system serves worst.

That instinct shows up in Delfina's go-to-market. The company did not start with a boutique concierge product for people who already have great care. It leaned into Medicaid and underserved populations - the hardest, highest-need cases - and built partnerships with community health providers like HOPE Clinic in Houston, Legacy Community Health, and Lone Star Circle of Care in Austin, plus health plans such as South Country Health Alliance. If you can move the numbers for the patients everyone else finds hardest to reach, the argument goes, the rest of the market is comparatively easy.