Knit Health

AI system trained on clinician decision-making to improve patient flow, triage, and operational workflows for health systems.

Website: https://www.knit.health/

Cover Block

Open sources

Field Detail
Company Knit Health
Tagline AI system trained on clinician decision-making to improve patient flow, triage, and operational workflows for health systems [Business Wire, May 2026]
Headquarters San Francisco, CA [ZoomInfo.com, 2026]
Founded 2026 [Business Wire, May 2026]
Stage Seed [Business Wire, May 2026]
Business model B2B
Industry Healthtech
Technology AI / Machine Learning
Geography North America
Growth profile Venture Scale
Founding team Academic spinout [University of California, Berkeley School of Public Health, July 2026]
Funding label Seed
Total disclosed funding $11.6 million [Business Wire, May 2026]

Links

Open sources

What an Investor Needs First

PUBLIC Knit Health is building an AI layer for health systems that aims to improve patient flow, triage, and care allocation by learning from real clinical decision-making, and it merits attention now because it emerged from stealth in May 2026 with an $11.6 million seed round and a data access story that is unusually central to the product claim [Business Wire, May 2026] [University of California, Berkeley School of Public Health, July 2026] [Knit Health, May 2026]. The company is described as a UC Berkeley spinout founded by Jon Kolstad, Maya Petersen, Jonas Knecht, and Ted Robertson, with public launch coverage tying the business to Berkeley research on clinical decisions rather than to a generic medical language model approach [University of California, Berkeley School of Public Health, July 2026] [LinkedIn, 2026].

The core product is what Knit calls a Large Clinical Behavior Model, positioned as a healthcare-native intelligence system that supports specialist routing, patient flow, and care team allocation; the differentiating claim is that the model is trained on clinician behavior in practice, using Truveta electronic medical record data spanning more than 130 million patients across 30 U.S. health systems [Business Wire, May 2026] [Unite.AI, May 2026] [Truveta]. Public materials point first to operational use cases such as predicting emergency department admissions, identifying patients ready for discharge, and finding candidates for hospital-at-home programs, which suggests the initial wedge is workflow and capacity management inside providers rather than broad clinical documentation or consumer health [Founderland AI, May 2026] [Knit Health, 2026].

The founding team's credibility appears strongest on domain knowledge and research depth. Public records identify Kolstad as a UC Berkeley professor and NBER research associate, while Berkeley coverage identifies Petersen as a professor in public health and computing; LinkedIn corroborates the four named co-founders, though public evidence on prior commercial scaling experience is thinner than the academic record [idcrawl.com, 2026] [University of California, Berkeley School of Public Health, July 2026] [LinkedIn, 2026].

On capitalization, Knit disclosed an undisclosed pre-seed led by Moxxie Ventures and an $11.6 million seed round co-led by Uncork Capital and Frist Cressey Ventures, with participation from Coalition Operators in the earlier financing [Business Wire, May 2026]. The business model is B2B, with sources describing the target customers as U.S. health systems and data providers, which is directionally attractive but still leaves commercial proof points, named customers, and deployment depth largely outside the public record today [Founderland AI, May 2026] [Business Wire, May 2026].

Over the next 12 to 18 months, the central questions are whether Knit can convert its data and research pedigree into referenceable provider deployments, whether the operational use cases produce measurable workflow gains inside health systems, and whether hiring in functions such as security, compliance, clinical informatics, and data engineering translates into enterprise readiness [TealHQ, 2026] [Business Wire, May 2026]. The public case is promising, but it still rests more on model thesis, founder quality, and early financing support than on independently verified customer traction.

Partially corroborated -- Core funding and founding claims are corroborated by Business Wire, Berkeley coverage, and LinkedIn, while several product and deployment details remain company-led or lightly corroborated.

Taxonomy Snapshot

Axis Value
Stage Seed
Business Model B2B
Industry / Vertical Healthtech
Technology Type AI / Machine Learning
Geography North America
Growth Profile Venture Scale
Founding Team Academic Spinout
Funding Seed, total disclosed ~$11.6M

Inside the Company

PUBLIC Knit Health surfaced publicly in May 2026 with a fairly specific claim on where it wants to sit inside hospital operations: not as a general clinical chatbot, but as an intelligence layer that helps route patients and clinical work across health systems [Knit Health, May 2026]. The company is headquartered in San Francisco, California, according to its company materials and third-party company profile coverage, and its public positioning ties the business closely to UC Berkeley research roots rather than to a prior operating business line [Knit Health, May 2026] [CB Insights, May 12, 2026].

The chronology is short but clear. Knit announced its launch from stealth on May 12, 2026 alongside an $11.6 million seed round, after an earlier undisclosed pre-seed backed by Moxxie Ventures and Coalition Operators [Business Wire, May 2026] [Knit Health, May 2026]. On its website and launch materials, the company describes its core product as a healthcare-native AI platform intended to improve patient flow, specialist routing, and care allocation for providers and healthcare innovators, which places the commercial focus on enterprise health system workflow rather than consumer care delivery [Knit Health] [Knit Health, May 2026].

Partially corroborated -- Confirmed primarily by company materials, with headquarters and launch details partially corroborated by CB Insights and Business Wire.

Under the Hood

Reported and inferred

Knit Health is not pitching a generic clinical copilot. The public materials describe a healthcare intelligence layer built to help health systems route patients, manage flow, and allocate clinical capacity in real time, with the key technical claim being that the model is trained on observed clinician decision-making rather than only text corpora or medical literature [Knit Health, May 2026] [University of California, Berkeley School of Public Health, July 2026] [Business Wire, May 2026]. On its homepage, the company says it provides "systemwide patient orchestration" and "continuously conducts clinical review for every patient seeking or receiving care, then guides each one to the right provider at the right time," language that frames the product as operational infrastructure for providers rather than a consumer-facing tool [Knit Health].

The core model is described publicly as a Large Clinical Behavior Model, or LCBM, trained on Truveta electronic medical record data spanning more than 130 million patients across 30 U.S. health systems [Unite.AI, May 2026] [Truveta] [HITConsultant.net, May 2026]. Reported use cases center on patient-flow operations: predicting emergency-department admissions, identifying patients ready for discharge, matching ambulatory patients to physicians, advanced practice providers, or virtual care, and forecasting transfers or hospital-at-home eligibility [Business Wire, May 2026] [Founderland AI, May 2026] [Knit Health]. Those claims suggest a workflow product that sits close to health-system operations and care coordination, though the public record does not establish performance benchmarks, integration depth, or named deployment references.

The technical stack itself is only lightly exposed in public. Berkeley's launch coverage says the company combines artificial intelligence with causal inference and reinforcement learning to model clinical decisions, while open roles in data engineering and security/compliance indicate the company is staffing for enterprise data handling and regulated deployment needs (inferred from job postings) [University of California, Berkeley School of Public Health, July 2026] [TealHQ, 2026]. That is directionally consistent with an enterprise healthtech product, but the available evidence still comes mostly from company and launch coverage rather than independent technical validation.

Partially corroborated -- Product positioning is corroborated across Knit Health, Business Wire, and UC Berkeley coverage, but capability detail and technical architecture remain only partially independently verified.

Market Research

PUBLIC

The market matters now because U.S. health systems are under pressure to move patients through scarce clinical capacity more efficiently, and Knit Health is positioning itself at the intersection of hospital operations, care coordination, and clinical decision support rather than in a single narrow software budget line [University of California, Berkeley School of Public Health, July 2026] [Business Wire, May 2026] [Knit Health, May 2026].

Public evidence does not support a direct TAM, SAM, or SOM for Knit specifically, so the cleanest read is through analogous markets. The company describes itself as selling to U.S. health systems and data providers, with initial use cases in emergency-department admissions, discharge readiness, hospital-at-home identification, transfers, specialist routing, and care-team allocation [Business Wire, May 2026] [Founderland AI, May 2026] [Knit Health, May 2026]. That places Knit across several existing spend categories at once: patient flow software, hospital operations analytics, care coordination, and AI-enabled clinical workflow infrastructure. The practical implication is that Knit may not enter the market through a single established budget owner, which can lengthen procurement, but it also widens the set of operational pain points the product can address if the workflow claims hold in deployment [Business Wire, May 2026] [University of California, Berkeley School of Public Health, July 2026].

Demand drivers are visible in the product framing itself. The company is not pitching ambient documentation or consumer navigation, two crowded AI lanes in healthcare, but rather tools that help hospitals decide who is likely to be admitted, who may be ready for discharge, who could be routed to lower-cost settings, and how scarce clinicians should be allocated in real time [Founderland AI, May 2026] [Knit Health, May 2026]. Those are operational problems with direct cost and throughput consequences for health systems. The underlying data story matters here: Knit says its model is trained on Truveta electronic medical record data spanning more than 130 million patients across 30 U.S. health systems, and Truveta separately describes its de-identified, daily refreshed dataset as covering more than 130 million patients with linked claims, mortality, and social drivers of health [Unite.AI, May 2026] [Truveta]. If accurate in production use, that breadth could make the platform relevant to both front-door triage and in-hospital flow, which are usually handled by separate operational teams.

The adjacent markets are at least as important as the core one. Knit can be read as a substitute for point solutions in bed management, transfer-center workflow, discharge planning, referral routing, and some forms of utilization management, while also overlapping with newer categories such as healthcare-native foundation models and enterprise clinical intelligence layers [Preqin, May 2026] [intelligence360] [Knit Health, May 2026]. That adjacency cuts both ways. It enlarges the theoretical market, but it also means incumbents in EHR workflow, care management, and hospital command-center software could claim partial feature overlap even if they are built on different technical assumptions. For investors, the more useful framing is not one market line but a stack of adjacent budgets that could converge if hospitals increasingly buy operational AI as a shared layer rather than as siloed tools [Business Wire, May 2026] [CB Insights, May 12, 2026].

Regulatory and macro forces support interest, though they also raise the diligence bar. Knit is hiring for a Head of Security & Compliance role, a signal that privacy, governance, and enterprise controls are not side issues in the sales motion [TealHQ, 2026]. The use of de-identified EHR data from a named source is directionally helpful for trust and model-development credibility, but hospital buyers will still care about explainability, workflow accountability, model drift, and how recommendations interact with clinician judgment [Truveta] [University of California, Berkeley School of Public Health, July 2026]. More broadly, the post-2024 healthcare AI market has rewarded products tied to measurable labor savings or throughput improvement more than open-ended intelligence claims, which makes Knit's focus on operational use cases commercially sensible, assuming it can translate research pedigree into procurement-grade evidence.

Market lens What public sources support Relevance to Knit
Hospital operations and patient flow Initial use cases include predicting emergency-department admissions, identifying discharge readiness, hospital-at-home candidates, and transfers [Founderland AI, May 2026] Suggests entry through capacity management and throughput workflows
Clinical intelligence infrastructure Knit describes a unified AI layer for specialist routing, patient flow, and care allocation [Knit Health, 2026] [Knit Health, May 2026] Suggests platform positioning beyond a single point solution
Health-system data and model development Truveta says its data covers more than 130 million patients and includes linked claims, mortality, and social drivers [Truveta] Supports the scale argument behind model training and potential generalizability
Data providers as buyers or partners Public summaries say Knit sells to U.S.-based health systems and data providers [Business Wire, May 2026] [Founderland AI, May 2026] Expands the addressable buyer set beyond provider systems alone

The market picture is broad rather than tightly bounded. Public sources support a credible demand case around patient flow and care allocation, but they do not yet support a quantified market size specific to Knit's category, so any sizing claim at this stage should be treated as analogous rather than company-specific.

Partially corroborated -- Section relies on named public sources for use cases, buyers, and data foundation, but no third-party TAM/SAM/SOM figures were confirmed for Knit's exact category.

Competition and Substitutes

MIXED Knit Health is positioning itself less against a single named startup peer than against a stack of existing hospital operations tools, EHR workflows, and internal analytics teams that already influence patient routing and capacity decisions [Business Wire, May 2026] [Knit Health, May 2026].

The public record here is thinner than the category warrants, which matters because competitive position in hospital software is usually defined by procurement alternatives rather than by direct startup lookalikes. On the incumbent side, health systems already rely on EHR-native workflows and care coordination logic embedded in large clinical systems, while operational decisions around discharge, transfer, and specialty routing are often supplemented by point solutions, consulting-led analytics, or homegrown models built by health-system data teams [University of California, Berkeley School of Public Health, July 2026] [CB Insights, May 12, 2026]. On the challenger side, Knit is presenting itself as a healthcare-native intelligence layer trained on observed clinical decision-making rather than on text alone, with initial use cases around admissions prediction, discharge readiness, hospital-at-home identification, and transfer forecasting [Business Wire, May 2026] [Founderland AI, May 2026].

That creates a distinct wedge, at least in the public materials. The strongest edge visible today is the data and research position: Knit says its model is trained on Truveta electronic medical record data spanning more than 130 million patients across 30 U.S. health systems, and Berkeley's July 2026 coverage ties the company closely to an academic founding team with backgrounds in health economics, causal inference, and clinical research [Unite.AI, May 2026] [Truveta] [HITConsultant.net, May 2026] [University of California, Berkeley School of Public Health, July 2026]. If that training approach translates into measurably better workflow recommendations inside live hospital settings, the combination of specialized data access and domain talent could be durable. If it does not, the edge is more perishable, because hospitals can default back to incumbent systems or ask existing vendors to add adjacent AI features.

The company's exposure is also fairly clear from what has not yet been disclosed. No specific health-system customer names or deployment references were verifiable in the cited public sources, which leaves distribution strength unproven relative to established hospital IT vendors and even to services-heavy analytics providers that already sit inside provider organizations [Founderland AI, May 2026] [Business Wire, May 2026]. Knit is also hiring for security, compliance, clinical informatics, and data engineering roles, a sensible signal for an enterprise healthcare seller, but also an indication that core go-to-market and implementation capabilities are still being built in parallel with the product [TealHQ, 2026] [Founderland AI, May 2026].

The most plausible competitive scenario over the next 18 months is not a clean winner-take-all outcome, but a sorting process around proof of deployment. Truveta looks like the clearest winner if health systems continue to value large, de-identified longitudinal datasets as the base layer for clinical operations AI, because Knit's public differentiation depends in part on that data substrate being hard to replicate [Truveta] [Business Wire, May 2026]. Knit would be the loser if customer references remain undisclosed and the category consolidates around incumbent platforms adding workflow intelligence directly into existing systems, because hospitals tend to privilege integration and procurement simplicity when the ROI case is still being established [CB Insights, May 12, 2026] [University of California, Berkeley School of Public Health, July 2026].

Opportunity

PUBLIC The prize here is not a better point solution for hospital operations, but a chance to become a system-level intelligence layer that health systems use to decide where patients go, which clinicians see them, and how scarce clinical capacity gets allocated across the enterprise [Business Wire, May 2026] [Knit Health, May 2026] [University of California, Berkeley School of Public Health, July 2026].

That outcome is still early, but it is not purely theoretical. Knit launched publicly in May 2026 with an $11.6 million seed round co-led by Uncork Capital and Frist Cressey Ventures, after an earlier pre-seed led by Moxxie Ventures, which gives it enough early backing to hire into security, compliance, clinical informatics, and data engineering rather than remain a research project [Business Wire, May 2026] [TealHQ, 2026]. More important, the product thesis is unusually specific: the company says its Large Clinical Behavior Model is trained on Truveta electronic medical record data spanning more than 130 million patients across 30 U.S. health systems, and that it is aimed at specialist routing, patient flow, care allocation, and related operational decisions where health systems already feel economic pressure [Unite.AI, May 2026] [Truveta] [Knit Health, 2026]. If that model can move from interesting predictions to trusted workflow recommendations, Knit could become default infrastructure for clinical operations rather than another analytics layer sitting beside the EHR [University of California, Berkeley School of Public Health, July 2026] [Knit Health, May 2026].

The most credible paths to scale are visible even from the limited public record.

Scenario What happens Catalyst Why it's plausible
Patient-flow wedge to enterprise control layer Knit starts with high-urgency use cases such as emergency department admissions, discharge readiness, transfers, and hospital-at-home identification, then expands into broader systemwide orchestration across inpatient and ambulatory settings A successful early deployment that shows recommendations can update in real time from live clinical data and earn operator trust inside a health system [Knit Health, 2026] [Founderland AI, May 2026] The company already frames initial applications around patient flow and routing, and its public product language centers on continuously guiding each patient to the right provider at the right time [Founderland AI, May 2026] [Knit Health]
Become the intelligence layer for data-rich health systems Knit becomes the model layer that health systems and data partners use to translate large EHR datasets into operational actions, not just retrospective insights Deepening of the Truveta-linked data foundation into production-grade enterprise use cases [Business Wire, May 2026] [Truveta] Public materials repeatedly tie the product to Truveta data covering more than 130 million patients across 30 U.S. health systems, which is a meaningful starting asset if model performance generalizes in live settings [Business Wire, May 2026] [Unite.AI, May 2026]
Clinical decision infrastructure for new care models Knit becomes the routing and allocation engine behind newer delivery formats such as virtual care mixes and hospital-at-home programs Wider provider demand for matching patients across physicians, advanced practice providers, and virtual care resources as capacity constraints persist [Knit Health] [intelligence360] The company already positions matching across MDs, APPs, virtual care, and other resources as a core function, which gives it exposure to care settings that extend beyond a single hospital unit [Knit Health] [intelligence360]

What compounding looks like is straightforward in theory, even if unproven in practice. A hospital or health system first adopts Knit for one narrow workflow where the operational pain is visible, such as admissions prediction or discharge timing [Founderland AI, May 2026]. If the recommendations prove useful, the same intelligence layer can extend into adjacent decisions, because routing, transfers, discharge planning, specialist allocation, and ambulatory matching all draw on overlapping patient context and similar workflow infrastructure [Knit Health, 2026] [Preqin, May 2026].

The compounding mechanism is partly data, partly workflow position. Knit says its system continuously updates recommendations as new clinical data arrive, which means product value should improve as it becomes embedded deeper into live operational loops rather than used as a static reporting tool [Knit Health] [intelligence360]. The hiring signal matters here as well: a posted Head of Security & Compliance role suggests the company is already orienting around the controls required for enterprise healthcare adoption, a necessary condition if it wants to move from pilot logic to durable operating-system status inside health systems [TealHQ, 2026].

The size of the win is best framed conditionally, because the public record here does not support a market size model. If Knit becomes a trusted operating layer for patient flow and care allocation across major U.S. health systems, the company would look less like a narrow workflow vendor and more like strategic infrastructure sitting on top of clinical data and care operations. In that scenario, not a forecast, the outcome could support a multibillion-dollar enterprise value, because infrastructure-grade healthcare software with embedded workflow ownership has historically attracted outsized strategic and investor interest even before category maturity [Business Wire, May 2026] [University of California, Berkeley School of Public Health, July 2026]. The public evidence is not yet strong enough to anchor that upside to a named revenue multiple or peer market cap, so the upside case rests on category position: if Knit becomes the place where operational clinical decisions are computed and routed, the company could matter at health-system scale rather than department scale.

Partially corroborated -- This section relies on one strong independent financing source, one independent university profile, and company materials for product scope; the upside framing is evidence-based but still early and partly dependent on company-described capabilities.

Sources

Open sources

  1. [Business Wire, May 2026] Knit Health Launches with $11.6M Seed to Build Clinical Intelligence AI for Healthcare. | https://www.businesswire.com/news/home/20260512133959/en/Knit-Health-Launches-with-$11.6M-Seed-to-Build-Clinical-Intelligence-AI-for-Healthcare

  2. [University of California, Berkeley School of Public Health, July 2026] UC Berkeley researchers, entrepreneurs create first healthcare AI model trained to understand clinical decisions. | https://publichealth.berkeley.edu/articles/news/first-healthcare-ai-model

  3. [Knit Health, May 2026] Knit Health raises $11.6M for ‘fundamentally different’ approach to healthcare AI. | https://www.knit.health/blog/knit-health-raises-11-6m-for-fundamentally-different-approach-to-healthcare-ai

  4. [LinkedIn, 2026] Ted Robertson - San Francisco, California, United States | Professional Profile | LinkedIn | https://www.linkedin.com/in/tedr2020/

  5. [Unite.AI, May 2026] Knit Health emerges from stealth with $11.6M to build AI that thinks like doctors. | https://www.unite.ai/knit-health-emerges-from-stealth-with-11-6m-to-build-ai-that-thinks-like-doctors/

  6. [Truveta] Truveta Data | https://www.truveta.com/data/

  7. [idcrawl.com, 2026] Jonathan Kolstad, Berkeley and NBER profile. | https://idcrawl.com/jon-kolstad

  8. [TealHQ, 2026] Head of Security & Compliance | https://www.tealhq.com/job/head-of-security-compliance_7ea1a6040d23513cc54b83369b574a9c3c3f3

  9. [ZoomInfo.com, 2026] Knit Health - Overview, News & Similar companies | https://www.zoominfo.com/c/knit-health-technologies-inc/5000849294

  10. [CB Insights, May 12, 2026] Knit Health - Products, Competitors, Financials, Employees, Headquarters Locations | https://www.cbinsights.com/company/knit-health

  11. [Knit Health] Knit Health | https://www.knit.health/

  12. [Founderland AI, May 2026] Knit Health Raises $11.6M to Build Clinical AI From Doctor Behavior. | https://www.founderland.ai/articles/knit-health-raises-116m-to-build-clinical-ai-from-doctor-beh-mp5nyxr2

Articles about Knit Health

View on Startuply.vc