AI Companies in Healthcare: A Structured Landscape Overview

The healthcare AI company landscape is genuinely difficult to read clearly. Vendor marketing, analyst estimates with wide methodological variance, and a steady stream of acquisition announcements make it hard to distinguish companies that have cleared regulatory hurdles and deployed at scale from those still in research or pre-revenue stages.

This overview organizes the landscape by what companies actually do — the clinical or operational problem they address — and anchors each segment to verifiable facts: FDA clearance status, disclosed funding or public market status, and whether peer-reviewed evidence exists for their products. The goal is orientation, not ranking.

How This Landscape Is Organized

Healthcare AI companies cluster into a small number of functional domains. The boundaries are not perfectly clean — a radiology AI company may also sell workflow tools — but the primary application area is usually clear from product portfolios and FDA submission records.

• Medical imaging AI (radiology, pathology, ophthalmology)
• Ambient documentation and AI scribes
• Clinical decision support embedded in EHR workflows
• Drug discovery and genomics AI
• Revenue cycle and administrative AI with clinical touchpoints

FDA clearance status matters here because it is the most verifiable signal of product maturity for clinical deployment in the US. Roughly 75% of FDA-cleared AI/ML-enabled medical devices are concentrated in imaging specialties, which is why imaging AI companies dominate the cleared-product count. That concentration reflects the data availability and task structure of imaging — not necessarily that imaging AI is more clinically impactful than other domains.

Medical Imaging AI Companies

This is the most populated segment of the cleared-device landscape. Companies here build AI tools for radiology, pathology, and other imaging-dependent specialties. Most products are cleared via the 510(k) pathway and address specific detection or triage tasks — pulmonary nodule detection, intracranial hemorrhage flagging, diabetic retinopathy screening, and similar applications.

Aidoc

Aidoc is a privately held company focused on AI-powered triage and clinical coordination tools for radiology. The company holds multiple FDA clearances across neurological, pulmonary, and vascular imaging applications. Its platform is deployed at a significant number of US health systems and integrates with major PACS vendors. Aidoc has disclosed Series C and later funding rounds; the total disclosed funding is in the hundreds of millions of dollars range. Peer-reviewed evidence for specific Aidoc products exists in the literature, though external validation studies remain limited relative to the breadth of the cleared product portfolio.

Viz.ai

Viz.ai operates in the AI-powered care coordination space, with FDA-cleared products for stroke, pulmonary embolism, and aortic disease detection. The company's model routes imaging findings to the appropriate care team in near-real-time. It is privately held and has raised substantial venture funding. Several published studies examine its stroke notification workflows, with some prospective data on door-to-treatment time reductions — though study populations are often single-center or limited to specific health system configurations.

Paige

Paige focuses on computational pathology — AI applied to whole-slide imaging for cancer detection. It received the first FDA authorization for an AI-based pathology product using a De Novo pathway. The company has partnerships with major cancer centers and has published peer-reviewed performance data. It is privately held with disclosed venture backing. The computational pathology segment is considerably less mature in terms of cleared products and clinical deployment breadth than the radiology AI segment, making Paige a notable exception.

iCAD

iCAD is a publicly traded company (NASDAQ: ICAD) with a long history in mammography computer-aided detection. Its ProFound AI product line holds FDA clearances for mammography and tomosynthesis. iCAD is one of the few smaller publicly traded pure-play healthcare AI companies, which makes its financial disclosures more transparent than most private peers. Published studies on ProFound AI exist in peer-reviewed radiology literature, including some prospective data on reader performance.

Intelerad and Nanox (Nanox.AI)

Intelerad acquired Nanox.AI (formerly USARAD and then Nanox AI following Nano-X Imaging's acquisition of Zebra Medical Vision) in a corporate structure that has shifted several times. Zebra Medical Vision's original FDA-cleared algorithms for bone density, cardiovascular risk, and liver fat detection are now part of this portfolio. Tracking cleared products through these ownership changes requires checking FDA submission records directly, as marketing materials do not always accurately reflect what is currently authorized.

Ambient Documentation and AI Scribe Companies

Ambient AI documentation — where a microphone captures the clinical encounter and AI generates a structured note — has seen rapid deployment growth since 2023. This segment is notable because most ambient AI scribe products are not FDA-regulated as medical devices; they are classified as general software tools or administrative aids. That means the evidence standards and oversight structures are different from the imaging AI segment.

Nuance (Microsoft)

Nuance Communications was acquired by Microsoft in 2022 for approximately $19.7 billion. Its DAX (Dragon Ambient eXperience) product is the most widely deployed ambient AI documentation system in US healthcare, with documented integrations across Epic, Cerner, and other major EHR platforms. Nuance holds some FDA clearances for specific speech recognition products, but DAX Copilot — the generative AI-enhanced version — operates under a different regulatory framework. Published peer-reviewed evidence on DAX's impact on documentation time and physician satisfaction exists, though most studies are single-institution or vendor-facilitated.

Abridge

Abridge is a privately held ambient AI company that has secured notable health system partnerships, including a widely publicized deployment with UPMC. The company uses large language model technology to generate clinical notes from encounter audio. It is not FDA-cleared as a medical device. Peer-reviewed implementation data is limited as of mid-2026; most available evidence is from vendor-disclosed metrics or institutional press releases. The company has raised substantial Series B and C funding.

Suki

Suki is a voice-based AI assistant for clinical documentation, privately held with Series C funding disclosed. It integrates with multiple EHR systems and is positioned at the intersection of traditional voice dictation and ambient AI. Like most ambient documentation tools, it is not FDA-regulated as a medical device. Independent peer-reviewed studies are sparse.

Clinical Decision Support Companies

Clinical decision support (CDS) AI covers a wide range of applications: sepsis prediction, deterioration alerts, medication safety, and diagnostic support. The regulatory picture here is more complex than imaging AI, because many CDS tools claim exemption from FDA oversight under the 21st Century Cures Act's CDS software provisions — a classification that has been contested and clarified in ongoing FDA guidance.

Epic

Epic is not primarily an AI company, but it is one of the most consequential actors in clinical AI deployment. Epic's EHR platform is used by a large proportion of US hospitals, and its embedded AI models — including sepsis prediction, deterioration risk, and readmission risk — are deployed at scale by default in many health systems. Epic's proprietary models are trained on its own multi-site data. The company is privately held. Several published studies have critically evaluated Epic's sepsis model (the Epic Sepsis Model, or ESM), with findings raising questions about its positive predictive value in external populations.

Wolters Kluwer Health (UpToDate, Medi-Span)

Wolters Kluwer is a publicly traded Dutch information services company. Its health division operates UpToDate and several clinical decision support and medication safety products. AI features have been progressively integrated into these platforms. It is not a pure-play healthcare AI company, but its CDS products are embedded in a large number of clinical workflows globally.

Sepsis and Deterioration AI Vendors

Several companies focus specifically on early warning and sepsis prediction, including Dascena (acquired), Philips' Early Warning Scoring tools, and BioVitals (Sensium). The evidence base for sepsis AI tools is uneven: some products have prospective validation data, others rely on retrospective performance metrics from the development institution. External validation — testing on patient populations from different health systems — remains a gap for most products in this segment.

Large Health Systems and Platform Companies with AI Divisions

Several large technology and health IT companies have substantial healthcare AI portfolios that don't fit neatly into the startup categories above.

GE HealthCare and Siemens Healthineers hold among the largest portfolios of FDA-cleared AI devices of any companies in this space, largely because their AI tools are integrated into imaging hardware that already goes through FDA review. This is a structurally different path to clearance than a standalone software company seeking 510(k) authorization for a software-only product.

Drug Discovery and Genomics AI

Drug discovery AI operates largely outside the FDA device clearance framework — these companies are developing molecules or identifying targets, not deploying software as a medical device. The evidence standard here is clinical trial outcomes for the drugs that AI-assisted pipelines produce, which is a longer and different validation cycle.

Recursion Pharmaceuticals

Recursion is a publicly traded (NASDAQ: RXRX) company that uses AI and high-content imaging to identify drug candidates. It has a large proprietary dataset of cellular imaging and has entered partnerships with major pharmaceutical companies. No FDA-cleared medical devices; the company's products are drug candidates in clinical development, not deployed clinical AI tools.

Tempus AI

Tempus AI (NASDAQ: TEM) operates at the intersection of genomic sequencing, clinical data, and AI. It provides molecular profiling services for oncology and has expanded into AI-based clinical insights products. Tempus went public in 2024. It holds some FDA authorizations for specific genomic assay products. The company's AI analytics layer sits on top of a large proprietary dataset of de-identified clinical and molecular data from health system partnerships.

Illumina

Illumina is primarily a genomic sequencing hardware and reagents company (NASDAQ: ILMN), but its DRAGEN AI platform for secondary genomic analysis is relevant here. DRAGEN uses AI-accelerated algorithms for variant calling and genomic interpretation. Illumina has FDA clearances for specific DRAGEN-based assays. It is a large-cap public company with a different risk and scale profile than the venture-backed AI startups in this landscape.

Revenue Cycle and Administrative AI

Revenue cycle AI tools — prior authorization automation, medical coding assistance, claims denial prediction — occupy a gray zone in this landscape. They have clinical touchpoints (prior auth decisions affect care access) but are typically not regulated as medical devices. Several companies in this space have grown rapidly.

• Cohere Health — prior authorization AI, privately held, Series C stage. Focuses on reducing friction in prior auth workflows for payers and providers.
• Waystar — publicly traded revenue cycle management platform with AI-powered claims and denial management tools.
• Olive AI — formerly a prominent healthcare automation company; underwent significant restructuring and product narrowing. A cautionary example of the gap between funding and durable deployment.
• Availity — health information network with AI features for eligibility and claims processing; privately held.

How to Read This Landscape: Practical Signals

The volume of companies claiming to deploy AI in healthcare makes it hard to distinguish substantive actors from early-stage or marketing-forward ones. A few signals help.

No single signal is sufficient. A company can be publicly traded with multiple FDA clearances and still have products with limited real-world outcome evidence. Conversely, a well-funded private company may have strong peer-reviewed data but no FDA clearance because its product falls outside the device regulatory framework.

Acquisitions and Consolidation

The healthcare AI company landscape has been consolidating. Several notable acquisitions have moved products between companies in ways that affect how cleared devices are tracked and supported.

• Microsoft acquired Nuance Communications (2022, ~$19.7B) — the largest healthcare AI acquisition to date.
• Nano-X Imaging acquired Zebra Medical Vision (2021), subsequently restructured as Nanox.AI.
• Oracle acquired Cerner (2022, ~$28.3B) — primarily an EHR acquisition, but with significant implications for embedded AI deployment.
• Amazon acquired One Medical (2023, ~$3.9B) — primary care, not AI-specific, but reflects platform companies expanding into care delivery.
• Intelerad acquired Nanox.AI (2024) — imaging AI portfolio absorbed into a larger radiology IT company.

Acquisitions complicate product tracking. When a startup with FDA clearances is acquired, the cleared devices remain authorized under the original submission number, but the responsible party (the legal manufacturer in FDA terms) changes. Checking the FDA device database directly — rather than relying on either company's marketing materials — is the only reliable way to confirm current authorization status after an acquisition.

What This Landscape Does Not Include

The landscape described here represents companies with documented market presence as of mid-2026. Given the pace of funding rounds, acquisitions, and product launches, specific details — particularly funding stages and product counts — should be verified against primary sources for any decision-relevant use.