Healthcare AI: what are the top startups?

Which healthcare AI startup became the doctor app everyone talks about?

OpenEvidence is the clear answer here.

In physician-facing AI, it is currently in a different league because the company has three signals at once: massive clinician usage, extreme valuation momentum, and a distribution model built around doctor attention.

The latest numbers are unusually strong. In July 2025, OpenEvidence announced a $210 million Series B at a $3.5 billion valuation. By October 2025, reports put the company around $6 billion. In January 2026, it announced a $250 million round at a $12 billion valuation. That means the company appears to have moved from roughly unicorn status to $12 billion in about a year.

That kind of valuation jump can look crazy if we only see it as a clinical search product. It makes more sense if we see OpenEvidence as a new attention layer for doctors. The company said that in December 2025 alone, verified U.S. physicians and healthcare professionals used the product for about 18 million clinical consultations. It also said that more than 100 million Americans were treated in 2025 by a doctor using OpenEvidence.

The comparison is important. Abridge and Ambience sit inside the clinical visit. Hippocratic sits around patient communication. Aidoc sits inside imaging workflows. OpenEvidence sits one step earlier: when the doctor is deciding what to know, check, or confirm. That is a much more frequent behavior. A doctor may use an AI scribe during visits, but they can use a medical answer engine across many clinical questions in a day.

There is also a business-model difference. Reports around the January 2026 round cited more than $100 million in annualized revenue, while the product remains free for verified clinicians and monetizes partly through pharma advertising. If we take the reported $12 billion valuation and the reported $100 million annualized revenue at face value, that implies roughly 120x annualized revenue. That is a very rich multiple. But it also tells us what investors are really pricing: not just software revenue, but a possible control point in physician attention.

If you want more recent data on this point, please see our latest healthcare AI market report.

Which healthcare AI startup is actually winning the AI scribe war?

Abridge is still the leader in the AI scribe war.

However, the gap is no longer just about who writes the best note but rather about who can turn ambient documentation into a broader clinical workflow platform.

Abridge has the strongest leadership case because it combines enterprise adoption, valuation, and product expansion. In June 2025, it raised $300 million at a $5.3 billion valuation. That came after reports of a roughly $2.75 billion valuation only a few months earlier. The market was clearly rewarding Abridge not as another AI note-taker, but as the likely enterprise standard for clinical conversations.

The important point is that Abridge is trying to move beyond the note. Its June 2026 Nvidia partnership matters because it suggests the company wants to use its clinical conversation footprint to build more healthcare-specific AI capabilities. Put simply: Abridge is not only listening to the visit but actually trying to make the visit data useful for documentation, coding, quality, and clinical intelligence.

Ambience Healthcare is the best challenger because it starts from a broader workflow thesis. In July 2025, it raised $243 million in Series C funding. It also reports more than 40 health-system clients, including major names such as Cleveland Clinic and Memorial Hermann. Compared with Abridge, Ambience has done a better job describing the whole hospital workflow around the note: documentation, coding, CDI, and administrative cleanup.

Nabla is the adoption challenger. In June 2025, it raised $70 million in Series C funding and said more than 130 healthcare organizations and 85,000 clinicians were using its assistant. That clinician number is the reason Nabla deserves to be in the conversation. Abridge feels stronger at the enterprise-platform layer, but Nabla has a credible claim that it is spreading widely across actual users.

Freed is the bottom-up outlier. It is not competing with Abridge on giant health-system contracts. Its wedge is independent clinicians who want something simple and cheap. In April 2025, Freed announced a $30 million Series A, and the company said it had passed 17,000 paying clinicians. Sacra estimated around $19 million ARR in March 2025. That is smaller than Abridge, but much cleaner as a product-led SMB adoption story.

Which healthcare AI startup is closest to replacing repetitive clinical labor?

Hippocratic AI is the standout.

Why? Because it is going after the repetitive patient-facing work that healthcare systems already struggle to staff.

That makes the company much more deployable than a vague “AI doctor” startup. Hippocratic AI focuses on non-diagnostic, non-prescribing tasks: post-discharge follow-up, medication reminders, patient education, scheduling support, and care coordination. These workflows are repetitive, high-volume, and expensive when done only by human staff.

The funding signal is strong. In January 2025, Hippocratic AI raised $141 million at a $1.64 billion valuation. In November 2025, it raised another $126 million at a $3.5 billion valuation, bringing total funding to $404 million. More importantly, the company says its agents have handled more than 180 million clinical interactions across more than 1,000 use cases with 60+ partners worldwide.

The comparison with OpenEvidence is useful. OpenEvidence is winning the doctor-knowledge layer. Hippocratic AI is winning the patient-communication layer. One helps clinicians think faster; the other helps healthcare organizations scale communication around the patient. They are both “healthcare AI,” but they touch totally different bottlenecks.

The comparison with consumer “AI doctor” apps is even more important. Many consumer health chatbots sound exciting but run into trust, liability, escalation, and reimbursement problems. Hippocratic AI avoids part of that trap by staying around the doctor, nurse, or health system rather than pretending to be the clinical authority itself.

The June 2025 Universal Health Services deployment for post-discharge engagement is a good example of why this wedge works. Hospitals do not need a science-fiction doctor bot but reliable follow-up at scale, especially after discharge, when missed instructions and poor engagement can create real costs.

All things considered, Hippocratic AI is the strongest patient-agent startup because it picked the right job.

If you want more recent data on this point, please see our latest healthcare AI market report.

Which healthcare AI startups are actually fixing the admin mess?

Tennr, CodaMetrix, Anterior, AKASA, and Arbiter are the names to watch.

Tennr is winning referrals, CodaMetrix is winning coding automation, Anterior is emerging on payer workflows, AKASA is the older RCM specialist, and Arbiter is the aggressive newcomer.

Tennr has the freshest momentum in referral automation.

In June 2025, it raised $101 million and launched Tennr Network to give providers more visibility into patient referrals. That sounds narrow until we remember how healthcare actually works. Referrals are still buried in faxes, PDFs, missing fields, manual phone calls, and unclear handoffs. If Tennr can turn that chaos into a real network, it is attacking one of the most painful administrative choke points in healthcare.

CodaMetrix is more mature and more deeply embedded in medical coding.

In June 2025, the company said its customers represented 220 hospitals across 27 states and $180 billion in net patient revenue, including 9 of the 20 U.S. News & World Report Honor Roll institutions. That is a very different proof point from Tennr’s. Tennr looks hotter and earlier. CodaMetrix looks more institutionally installed.

Anterior is the payer-workflow startup that has become harder to ignore.

In 2026, it raised $40 million, bringing total funding to $64 million, to scale AI for health plan workflows and prior authorization. The reason Anterior matters is that healthcare AI cannot only automate the provider side. If hospitals get better AI for documentation and billing, payers will also need better AI for clinical review, prior authorization, and medical policy.

AKASA remains relevant, but its story feels less fresh than Tennr or Anterior.

It has sold into major hospitals such as Cleveland Clinic, Duke, and Johns Hopkins, according to investor commentary, and it remains one of the better-known AI revenue-cycle automation companies. But in a “who is hot right now?” ranking, it does not have the same recent acceleration signal.

Arbiter is the wildcard.

In November 2025, Business Insider reported that it raised $52 million in seed funding at a $400 million valuation. The company said it was supporting more than 1,000 clinicians and building an administrative “operating spine” for healthcare. That is still early, and we should not confuse a big seed round with proof of category leadership. But it is exactly the kind of signal that says a new company is trying to collapse several admin workflows into one layer.

Which healthcare AI startups still matter in medical imaging?

Aidoc is still the leader in imaging AI. Rad AI and Qure.ai are the two most credible challengers, but they are winning different arguments.

Aidoc has the clearest enterprise and regulatory case. In April 2026, it raised $150 million in Series E funding led by Goldman Sachs Alternatives, bringing total funding to more than $500 million. Axios also reported that the company had 31 FDA clearances and that CEO Elad Walach was targeting an IPO within three to five years.

This is an important point. Imaging AI is one of the few parts of healthcare AI where “regulatory proof” is not optional. Many AI startups can show demos. Imaging companies need clinical validation, FDA pathways, hospital integration, and enough trust for physicians to use the output. Aidoc has been building exactly that for years.

Rad AI is the more generative-AI-native challenger. In January 2025, it raised $60 million at a $525 million valuation. Its stronger angle is radiology productivity: reporting, follow-up, and workflow automation. That may be more valuable now than simply detecting another finding on an image. Radiology teams are overloaded, so the winner may be the company that saves radiologist time across the whole workflow.

Qure.ai has a different kind of strength. It is not just a U.S. hospital AI story.

In January 2026, it received a multi-million-dollar Gates Foundation grant to advance AI diagnostics for tuberculosis and pneumonia in under-resourced regions. That gives it a stronger global-health and infectious-disease positioning than most imaging AI startups.

Aidoc is first because it has the deepest regulatory and enterprise case. Rad AI is the productivity challenger because it is closer to the daily radiologist workflow. Qure.ai is the global deployment specialist.

Lately, imaging AI looks less like the hottest frontier and more like the most mature healthcare AI category. That is not a bad thing. It means the best companies here are being judged on deployment, clearances, and clinical workflow, not just model performance.

If you want more recent data on this point, please see our latest healthcare AI market report.

Which healthcare AI startup is making oncology less painfully manual?

Triomics is the most interesting emerging name right now in oncology AI.

It is not the biggest healthcare AI startup, but the signal is very clean: fresh funding, a very specific workflow, and real oncology-system relevance.

In May 2026, Triomics raised $22 million in Series B funding led by Battery Ventures. The company is focused on oncology workflows such as trial matching, pre-visit chart preparation, and cancer data abstraction.

That sounds operational, but in oncology, operational friction is a real clinical problem. Cancer records are long, fragmented, and full of unstructured details that affect treatment and trial eligibility.

This is where Triomics compares well against broader clinical AI tools. A generic scribe may summarize a visit. Triomics is closer to the messy oncology data layer: extracting what matters from charts, matching patients to trials, and helping cancer teams prepare faster. That is a narrower market, but a much sharper product wedge.

The customer and partner signals also matter. Triomics said strategic backers included Oncology Ventures and Precision Health Informatics, a wholly owned subsidiary of Texas Oncology. Reports also pointed to expansion with Yale New Haven Health into AI-enabled cancer registry abstraction and reporting. Those signals are more relevant than a generic “health-system customer” claim because they map directly to oncology workflows.

Atropos Health still matters in real-world evidence. It is not as fresh as Triomics in this specific “who is hot now?” frame, but its GENEVA OS and ChatRWD products speak to a real need: faster evidence generation from clinical data. Unlearn also remains credible in clinical-trial digital twins. Its $50 million Series C in 2024 showed that investors still believe AI can change trial design.

But if we are being strict about recent evidence, Triomics ranks first in this category. Atropos and Unlearn are important, but their strongest public signals are older or less oncology-specific.

Triomics gives us a May 2026 financing event, a painful oncology workflow, and proof that oncology-specific buyers are paying attention.

That is enough to make it the emerging oncology AI startup to watch now.

Which healthcare AI startups are most credible in drug discovery?

Xaira, Isomorphic Labs, Chai Discovery, Genesis Therapeutics, and Formation Bio are, today, the core names.

Xaira is the capital-scale moonshot. It launched in 2024 with more than $1 billion in initial funding, which is almost absurdly large for a startup launch. The reason Xaira belongs at the top is not that it has already proven better medicines. It has not. The reason is that it has the capital, scientific network, and ambition to build a full AI-native drug discovery platform from the ground up.

Isomorphic Labs is even more powerful, but it is a special case because it comes from Alphabet’s DeepMind orbit. In May 2026, reports said it raised $2.1 billion in Series B funding led by Thrive Capital. Its pharma partnerships with Eli Lilly and Novartis make it one of the most important AI drug discovery companies in the world. Still, we should treat it differently from a normal independent startup because it is backed by Alphabet’s infrastructure and reputation.

Chai Discovery is the emerging model-platform breakout. In December 2025, it raised $130 million at a $1.3 billion valuation, co-led by General Catalyst and Oak HC/FT. The company had launched Chai-2, a multimodal model for medical research and biologics design. Compared with Xaira, Chai is earlier and smaller. Compared with most AI biology startups, it has a stronger recent signal that investors believe its model layer can matter.

Genesis Therapeutics has the best fresh pharma-validation signal. In May 2026, Incyte expanded its collaboration with Genesis and committed $120 million upfront, split between $80 million in cash and $40 million in equity. The deal added at least five collaboration targets and included research funding for AI model training and inference. That is more convincing than a vague partnership announcement because there is real upfront money attached.

Formation Bio is different again. It is more about AI-enabled drug development than pure AI model discovery. Its $372 million Series D in 2024 remains a serious signal, especially with investors like Andreessen Horowitz, Sanofi, Sequoia, and Thrive. But compared with Chai’s December 2025 raise or Genesis’s May 2026 Incyte expansion, the freshest public momentum is less sharp.

If you want more recent data on this point, please see our latest healthcare AI market report.

Which healthcare AI startups are changing care delivery, not just software workflows?

Twin Health and Cadence are the strongest names here.

Knownwell is worth watching, but it is still more of an AI-enabled obesity care company than a proven AI care-delivery platform.

Twin Health is the strongest AI-specific care-delivery story. In August 2025, it raised $53 million to expand its AI digital twin platform for metabolic health. The company uses connected-device and clinical data to create a real-time model of each member, with a focus on diabetes, obesity, and metabolic disease.

What makes Twin more interesting than a normal digital-health coaching company is the outcomes evidence. Reports around the financing pointed to Cleveland Clinic-led randomized clinical trial results published in NEJM Catalyst, showing meaningful improvement in glycemic control and medication reduction. That matters because care-delivery startups often talk about engagement, while Twin can point to clinical outcomes.

Cadence is the scale leader in proactive chronic care. Its 2025 outcomes report said it helped more than 74,000 seniors manage chronic conditions and recover safely at home. In July 2025, it launched an Advanced Primary Care Management business under a CMS pathway, using AI and connected devices to help health systems manage patients outside the clinic. Business Insider also reported that Cadence had 100% year-over-year revenue growth in 2025 and was on track to surpass $100 million ARR by 2026.

The comparison is quite interesting. Twin has the stronger AI-native story because its core idea is a digital twin for metabolic care. Cadence has the stronger operational scale story because it is building chronic-care infrastructure with health systems. Twin feels more clinically differentiated. Cadence feels more commercially scaled.

Knownwell is earlier, but obesity care is too important to ignore. In October 2025, it raised $25 million led by CVS Health Ventures to expand clinics, virtual care, AI, clinical decision support, and value-based care infrastructure. The signal is real, especially with GLP-1 demand reshaping obesity care. But Knownwell’s AI proof is less central than Twin’s, and its scale proof is less developed than Cadence’s.

So it looks like Twin is the stronger AI-care model, while Cadence is the stronger scaled-care platform. That distinction matters. One is proving AI can personalize metabolic care. The other is proving technology-enabled care operations can reach a large chronic-care population.

Which healthcare AI categories are already getting less exciting?

Generic AI scribes are the first category to be careful with.

The market is still big, but basic note generation is no longer enough. These days, the winners need to move into coding, CDI, EHR actions, revenue cycle, or clinical intelligence. That is why Abridge, Ambience, Nabla, and Freed still matter, while many smaller “AI note-taker” startups are starting to blur together.

Consumer “AI doctor” apps are another weak area unless they have a serious clinical workflow behind them.

The problem is not that consumers do not want medical answers. They do. The problem is liability, trust, escalation, reimbursement, and integration with real care. OpenEvidence works because it targets verified clinicians. Hippocratic AI works because it stays in non-diagnostic patient communication. A consumer chatbot that simply says “ask me anything about your health” is much harder to rank as a serious startup.

Narrow diagnostics tools without workflow distribution also feel less exciting than they did a few years ago.

Aidoc still matters because it has regulatory clearances, hospital distribution, and a platform story. Rad AI still matters because it helps radiologists work faster. Qure.ai still matters because it has global deployment logic. A startup that only detects one condition, without owning the workflow around that detection, has a much harder path now.

At the end of the day, healthcare AI is becoming less forgiving. The market is moving away from impressive demos and toward workflow ownership. The companies that still look exciting are the ones that can show where they sit, who uses them, what gets automated, and why the buyer would keep paying.

Which healthcare AI startups are actually the top ones overall?

OpenEvidence, Abridge, Hippocratic AI, Aidoc, Ambience Healthcare, Xaira, Isomorphic Labs, Chai Discovery, Genesis Therapeutics, Tennr, CodaMetrix, Triomics, Anterior, Nabla, Twin Health, and Cadence are the strongest names overall.

OpenEvidence is the top physician-knowledge company because it has the most explosive combination of usage and valuation momentum. Abridge is the top clinical workflow company because it is turning the visit conversation into an enterprise platform. Hippocratic AI is the top patient-agent company because it picked a deployable labor wedge instead of pretending to be a doctor. Aidoc is the top imaging AI company because it has the regulatory and enterprise maturity that most healthcare AI startups still lack.

Ambience and Nabla are the serious Abridge challengers. Ambience is stronger on the full workflow story. Nabla is stronger on visible clinician adoption. Tennr and CodaMetrix show that healthcare admin AI may become one of the biggest infrastructure markets. Anterior is the emerging payer-side company to watch because AI automation cannot stop at the hospital door.

Triomics is the freshest oncology workflow startup. Chai Discovery and Genesis Therapeutics are the emerging AI drug discovery names with the most interesting recent signals: Chai through its December 2025 financing and model-platform positioning, Genesis through its May 2026 Incyte expansion. Xaira and Isomorphic are harder to compare with normal startups because their capital scale is so extreme, but they obviously belong in any serious healthcare AI ranking.

If we narrow the “top emerging” list, the names that stand out most are Triomics, Anterior, Tennr, Chai Discovery, Genesis Therapeutics, Arbiter, Nabla, and Freed. They are not all equally mature, but each has a recent signal that says something has changed lately: a new financing, a major partner, a measurable user base, a clearer category wedge, or a faster path into a painful workflow.

Finally, everything considered together, the top healthcare AI startups are not the ones with the broadest AI story but the ones with the sharpest workflow control. OpenEvidence owns clinical knowledge moments. Abridge owns clinical conversations. Hippocratic owns patient outreach. Aidoc owns imaging workflow. Tennr, CodaMetrix, and Anterior are moving into the administrative core. Triomics is going after oncology complexity. Chai, Genesis, Xaira, and Isomorphic are pushing the frontier of AI-native drug discovery. That is the real map of healthcare AI leadership right now.

If you want more recent data on this point, please see our latest healthcare AI market report.

Category	Startups selected and why
Physician medical knowledge	OpenEvidence leads because it combines the strongest recent usage signal, the fastest valuation jump, and a position at the doctor-attention layer rather than inside only one hospital workflow.
Ambient clinical documentation	Abridge leads because it has the strongest enterprise-platform signal; Ambience is the best full-workflow challenger; Nabla has the clearest broad clinician-adoption signal; Freed is the independent-practice breakout.
Patient-facing healthcare agents	Hippocratic AI leads because it focuses on non-diagnostic, repetitive clinical labor, where deployment is more realistic than consumer “AI doctor” positioning.
Healthcare admin automation	CodaMetrix has the deepest coding footprint; Tennr has the freshest referral-automation momentum; Anterior is the emerging payer-workflow bet; Arbiter is the high-risk new entrant.
Medical imaging AI	Aidoc leads through regulatory clearances, funding, and enterprise maturity; Rad AI is the radiology productivity challenger; Qure.ai is the global infectious-disease diagnostics specialist.
Oncology and clinical trials	Triomics is the freshest emerging oncology AI name because its May 2026 funding, customer logic, and workflow focus all point to a real category wedge.
AI-native drug discovery	Xaira and Isomorphic are the capital-scale giants; Chai is the emerging model-platform breakout; Genesis has the strongest recent pharma-validation signal; Formation Bio remains the AI-enabled development player.
AI-enabled care delivery	Twin Health has the stronger AI-native clinical model; Cadence has the stronger scaled chronic-care platform; Knownwell is worth watching in obesity care.

Who is the author of this content?

NEW MARKET PITCH TEAM

We track new markets so founders and investors can move faster

We build living "market pitch" documents for emerging markets: AI, synthetic biology, new proteins, and more. Instead of outdated PDFs or hallucinated LLM answers, our clients get a clean, visual, always-updated view of what's really happening: key players, deals, regulations, and signals that matter. Learn more about us.