Knit Health said Monday it has raised $35 million in a Series A round led by NEA, with participation from Frist Cressey Ventures, Town Hall Ventures and SignalFire. Existing backers Uncork, Moxxie, The Aligned Fund and Coalition Operators also took part.
The San Francisco company, which calls itself a healthcare AI business building a foundational model for predicting and improving patient care, announced the round on November 16, 2026. It has now raised $46.6 million in total, according to a company-issued Business Wire release reproduced by AOL. Knit declined to disclose a valuation, and PitchBook's public funding table leaves the post-money figure blank.
Moving from research to deployment
The money marks a transition. Knit is pushing its Med-BEAM model out of research and into commercial use, and the company said the proceeds will fund a rollout across health systems and other healthcare organizations.
The announcement includes a quote from Blake Wu, a partner at NEA. The company itself offered a bolder line: "Med-BEAM represents a fundamentally new clinical intelligence infrastructure."
What Med-BEAM is meant to do

Knit said the model was trained on billions of healthcare decisions and hundreds of millions of longitudinal patient records, meaning records that follow individuals over time. Those figures are the company's own, and Knit did not supply exact counts.
Chief executive Jonathan Kolstad said the first applications will be patient navigation, specialist routing, scheduling and hospital operations. He also described a longer-term plan to let outside organizations and developers build their own applications on top of the model.
Knit frames large language models and so-called world models as complements rather than rivals. In its telling, the language models provide the interface, while Med-BEAM models the clinical environment underneath.
The benchmark claims

In a separate post the same day, Knit reported that Med-BEAM delivered 2.48 to 4.56 times higher average precision than large language models, with inference 108 to 204 times faster and estimated costs 37 to 257 times lower. The company said the language-model costs reflected actual benchmark usage at published rates, whereas Med-BEAM's figure was its estimated compute expense multiplied by three.
These are company evaluations, not independent replications. Knit said it plans broader benchmarking across tasks, models and settings. Notably, neither the funding announcement nor the action-prediction benchmark shows improved patient outcomes from prospective deployment, the kind of evidence industry observers often note health systems look for before committing.
A short funding history
Knit emerged from stealth on May 12, 2026, announcing $11.6 million in seed funding co-led by Uncork Capital and Frist Cressey Ventures. An earlier pre-seed round was led by Moxxie Ventures, with Coalition Operators participating. It is unclear how large that pre-seed was, or whether it is counted within the $11.6 million.
The company says it was founded in 2025 by scientists from UC Berkeley. TipRanks listed 21 employees as of November 16, a figure Knit has not confirmed.
Neighbors and competitors
Knit is not alone among investors' healthcare AI bets. NEA and Frist Cressey also back Basalt Health, which announced a $20 million Series A on September 24, 2026, for AI-supported post-acute admissions and patient transitions. Tennr, which automates referral intake and routing, raised a $101 million Series C led by IVP, Fierce Healthcare reported on June 18, 2025.
What comes next

Knit said it will expand its scientific and engineering teams and bring the technology into additional clinical settings. Its job board offers a clue about another priority: a security role listed at roughly $200,000 to $265,000 a year. The posting says the company has relied on an outsourced virtual chief information security officer while it builds security leadership in-house, a modest but telling detail for a startup that hopes to handle sensitive patient records at scale.
