Billions in Play: Healthcare AI’s Race for Market Dominance
Author:
Table of Contents
- Introduction
- Bullish Buying: Healthcare’s Accelerating Rate of AI Adoption
- A Winner Take All Market: Investor Shift to Selective Value Concentration
- The Winning Playbook: Key Attributes of Leading Healthcare AI Companies
- Choppy Seas Ahead: Threats and Opportunities for Leading Companies
- Conclusion
- Strategic Imperatives for AI Buyers and Builders
Introduction
Much has changed since we published our perspectives on the healthcare industry’s accelerating adoption of artificial intelligence (AI), last year. Since then, breakthroughs in model architectures and training techniques have dramatically boosted AI performance while reducing compute and data requirements. At the same time, the proliferation of specialized AI hardware, managed cloud services, and open-source model hubs has democratized access, enabling rapid deployment of generative and agentic systems across industries. It’s therefore not surprising that AI usage and adoption are skyrocketing as model performance and user value improve just as dramatically as inference costs fall. Simply put, AI is becoming an indispensable part of our personal and professional lives.
This has profound implications for the future of work in healthcare, where AI is playing an integral role as an equalizer, extender, and enhancer of the workforce. The results so far are significant and tangible, fueling accelerated adoption. Consider the fact that two-thirds of physicians report currently using healthcare AI products, a 78% year-over-year usage increase. This rate of uptake is remarkable, especially compared to the rate of electronic health records (EHR) adoption over the last decade. There is a meaningful level of product-led “pull” driving this cycle of technology adoption, a stark contrast to the “push” that drove many previous adoption cycles.
As a result, healthcare enterprise leaders now view AI as much more than just an enabler of operational productivity. In fact, enterprise AI strategies are informed by recruitment and retainment priorities. In response, healthcare leaders are universally increasing their budgetary allocations to AI, concentrating on a handful of distinct categories of solutions where there is clinician and patient demand as well as proven ROI to rationalize budget expansion. External partnerships are also proving more successful than internal AI builds, creating strong tailwinds particularly for emerging AI solutions.
This has attracted a crowded cohort of emerging AI companies, across categories. Whereas just a year ago, the leading AI companies in each category were raising capital simply to capture market share, today many are seeking to break out of the pack by growing horizontally and bridging categorical boundaries. Meanwhile, incumbents are protecting moats by touting their own, emerging AI capabilities. And quietly, new upstarts are tapping into the forefront of technological progress to develop their own leads. How boundaries shift as leading companies strategically maneuver will shape the next few years of healthcare AI.
In this analysis we:
- Analyze market signals to gauge where we are in the healthcare AI adoption cycle
- Dissect funding data to understand where AI is creating durable value in healthcare
- Share our perspectives on why certain categories of healthcare are currently more primed for AI-led disruption than others
- Study the leading AI categories and companies to highlight key factors driving their success and threats they may face
- Offer suggestions for buyers scaling their AI partnerships, and builders looking to strengthen their moats
Bullish Buying: Healthcare’s Accelerating Rate of AI Adoption
The healthcare sector is emerging as a leading adopter of AI, bucking its longstanding reputation as a laggard of novel technology adoption. The charts below, sourced from the US Census Bureau and included in BOND Capital’s most recent AI Trends report, suggest that the healthcare industry’s rate of AI adoption is the fourth fastest across major sectors of the US economy in 2025. Since we started tracking this metric a year ago, adoption rates in healthcare have doubled.
As mentioned previously, this phenomenon is largely being driven by fundamentally different enterprise buying behavior. Earlier this year, we surveyed over 30 senior executives across our strategic healthcare partner base to understand their technology and AI priorities. The composition of the surveyed group and summary of survey insights are included in Figure 2, below.
On average, leaders across all sectors devote around 10% of their information technology (IT) budget to AI. Over half of the healthcare executives we surveyed allocate an even larger part of their IT budgets to AI, signaling a sustained, rather than fleeting, commitment to the technology. It is worth noting that of the four leaders who signaled that they were devoting more than 25% of their IT budgets to AI, three were hospital executives. As discussed in more detail later, hospitals seem to be outpacing the rest of the healthcare sector in AI adoption.
This pace of AI adoption is expected to persist, given nearly 83% of the healthcare executives we surveyed planned to increase their AI and technology budgets over the next 12 months. Interestingly, 75% of these leaders are not defaulting to incumbent solutions for AI needs, but instead seeking best-of-breed emerging technology solution partners, underscoring a particular enthusiasm about emerging AI products.
This explains the frenzied levels of venture capital funding we are seeing in emerging healthcare AI products, shown in Figure 3 below. Through the first half of this year, nearly 58% of healthcare deals were in AI companies, a record pace. Importantly, unlike previous AI hype cycles, it appears that we have also broken out of the typical funding plateau of “testing and learning” that seems to accompany each major AI advancement, and into accelerated adoption. This sharp spike in funding is undoubtedly a function of investors reacting to strong commercial buying signals buried in unprecedented growth rates and pipelines, but also the fact that AI is becoming a standard feature in a broader swath of healthcare product offerings.
A Winner Take-All Market: Investor Shift to Selective Value Concentration
Unsurprisingly, the demand for high-quality healthcare AI companies that can rapidly capture growing budgets is exceeding historical supply, which shows up in valuation data. The charts in Figure 4 below contain valuation data for AI companies selling into hospitals, health plans and life sciences / pharmaceutical companies in 2024 vs. 10-Year Averages.
The average valuation of a healthcare AI company raising capital in 2024 was 50% higher than the healthcare industry average last year. The contrast is most stark in early-stage funding data, but also in companies selling into hospitals, where again, it appears that early-stage AI solutions seem to have taken a faster foothold. This phenomenon and its derivative impact have been referenced in several health insurer earnings calls throughout the year.
Another phenomenon we have been tracking is one of capital concentration. While the charts above seem to imply that later stage AI valuations have dramatically corrected, the underlying data actually shows that a small handful of companies are breaking out as winners in their respective categories. Further, capital and value are concentrating on them.
In fact, at least ten healthcare AI companies have achieved valuations greater than $1 billion over the last year, and at least five multi-billion-dollar healthcare AI exits have been announced so far this year, which is important, because it suggests that there are compelling liquidity opportunities emerging for this first wave of AI products. Firstly, this is a critical signal that AI solutions are creating durable value. Secondly, exits of this magnitude sustain larger capital inflows into the healthcare AI category. But thirdly, the magnitude of these outcomes alone are quite significant, given the average “successful” healthcare and digital health exit size has historically been $100 million or more. Interestingly, five of the at least seven total liquidity events that have been announced this year exceeding $1 billion in transaction value, have been for healthcare AI platforms.
As alluded to earlier, this value creation has been relatively concentrated. We refreshed the analysis that we created last year by incorporating 2024 venture capital funding data and what we continue to observe is that AI is creating outsized value in revenue cycle management and claims operations, patient engagement and contact center and clinical operations categories. Collectively, we refer to these categories as “administrative” in nature.
As highlighted in the chart below, AI companies targeting administrative functions have consistently delivered superior investor returns, commanding valuation premiums (a proxy for commercial revenue) around 30% higher than other categorical segments. But importantly, these categories are also sustaining the first multi-billion-dollar valuations (and liquidity outcomes) in healthcare AI. Six of the highest-profile healthcare AI unicorns minted so far this year, including our portfolio company, Smarter Technologies (fka SmarterDx), primarily address administrative inefficiencies across healthcare. Clearly, these categories also harbor significant appeal to buyers and upstream funding partners.
While administrative operations remain ripe for AI-enabled disruption, most venture capital funding continues to go towards clinical value propositions, a continuation of a 10-year phenomenon that we noted in our report last year. This reflects a shared belief that the largest future opportunity for AI-enabled value creation in healthcare will be in clinical care delivery and clinical decision support.
Parts of this future are steadily becoming a reality. Today, the average physician’s panel size is between 2,300 to 2,900 patients. Some hospital leaders that we have spoken to have developed internal mandates to increase panel sizes to 10,000 patients in the coming years, a feat that would require technology and AI, given existing labor constraints. By streamlining the diagnostic journey and reducing repeat testing, imaging AI platforms like Heartflow seem to have a promising role to play in this future. We are also seeing the emergence of AI-native care delivery models, like Jimini Health or Counsel Health, that promise to extend clinical capacity by sequencing AI and human clinician mediated care. Some of these care delivery models natively integrate AI to automate administrative tasks like billing, calls, and annotation, further boosting clinical productivity.
Despite the promise of clinical care delivery AI models, this sub-category of clinical solutions has a higher threshold of adoption and therefore more left to prove in order to achieve ubiquity. The simple fact that clinical AI products, broadly, continue to exhibit lower graduation rates from seed to later stage financing rounds (compared to administrative AI solutions), highlights this. In fact, clinical AI company valuations, on average, were about 25% lower than companies focused on administrative operations in 2024. An important unlock for many clinical care delivery AI companies is consistent demonstration of quality and safety, at a universally accepted standard. This will gate their ability to achieve payer coverage, at scale. Regulatory complexities, lengthy development timelines, reimbursement variability, and clinical integration complexity are other ongoing hurdles.
Some of the dichotomy between administrative AI uptake rates and clinical AI uptake rates can be explained by the activities that comprise each of these categories. Anthropic recently published an economic index, analyzing the occupations and skills that their Claude.ai model is being used for. In their analysis, below, two things jump out.
Firstly, industries requiring high levels of training, nuanced human tact and correlating to higher wages, like many parts of the healthcare industry, saw a much lower percentage of their jobs being impacted by AI. Secondly, today’s leading models excel at very specific activities like information summarization and reasoning (i.e., clinical chart review, medical necessity research), listening (i.e., scribing and automated speech recognition, call automation), writing and content creation (i.e., appeal and denial letter automation, systems biology modeling), and coding and programming (i.e., medical billing and coding, clinical documentation improvement). However, Claude.ai was more sparsely used for scientific research and technical analysis (i.e., clinical research and decision making), complex human interaction and people management (i.e., live patient care) or physical interventions and activities (i.e., surgical procedures).
This maps fairly consistently to the categories of healthcare where we have historically seen high versus low AI impact. Categories that consist of a higher volume of occupational skills that can easily be automated by AI are easier automation targets. Importantly, large language models (LLMs) are distinguished in their ability to structure previously unstructured data. Therefore, we are seeing the highest incremental gains in performance and automation within previously unstructured parts of the aforementioned workflows.
This skills distribution also suggests interesting potential for a few emerging categories of healthcare AI, most notably, data-enabled clinical decision support activities. This is one sub-category of clinical AI that is less constrained by some of the hurdles mentioned above and therefore has benefitted from higher uptake. Perhaps the best representation of companies harnessing this potential today are OpenEvidence or Layer Health (a Flare Capital portfolio company), which use LLM advancements to convert previously unstructured medical literature or chart data into queryable formats. These products can comprehend new patterns and derive novel clinical and financial inferences from the newly structured data, setting them up to guide fundamentally better clinical decision making. By both streamlining search and recall work at higher accuracy rates,and allowing clinicians to access insights at their disposal (vs. pushing them into live clinical workflows), clinical decision supports products like these are driving unprecedented adoption. There are new personal applications of this functionality beginning to emerge as well.
We believe these are just a few of many emerging opportunities as clinicians and patients develop more comfort with the latest AI tools.
The Winning Playbook: Key Attributes of Leading Healthcare AI Companies
All of these adoption and budgeting patterns are shaping a dynamic of capital concentration. Today, six categories of healthcare seem to be capturing an outsized proportion of current enterprise AI budgets — including, AI physician assistants, billing and coding, care management, member and patient services, network and clinical operations, as well as radiology and imaging solutions.
These categories share several unique market attributes that make them popular targets for AI-enabled transformation:
- First, each of these categories have high technical automation potential given they are comprised of a high volume of manual activities that require the occupational skills discussed above
- Second, because these activities are more manual, they result in higher amounts of data lossiness (i.e., information discussed but not captured on a phone call)
- Third, that lossiness creates friction or erodes value for a core healthcare stakeholder, such as a clinician or patient
- Lastly, these categories support broad product surface areas, meaning there are lots of adjacent inefficiencies that a product can naturally expand to beyond its entry wedge (i.e., physician assistant scribes expanding into billing and coding workflows, or call scheduling solutions expanding into patient throughput activities)
Emerging products in these categories are targeting these inefficiencies by:
- Automating activities to drive productivity
- Capturing and structuring lost data, unstructured data or data exhaust to improve the effectiveness of these activities and inform new activities
- Creating a superior experience for a core healthcare user, usually by saving them time or helping them make faster decisions
- Evaluating the derivative impact of these activities on adjacent activities, to launch new products that extend their ROI
Naturally, all of the companies in these categories have earned product-market fit by addressing the inefficiencies discussed above. However, this is table stakes. What has distinguished category leaders from their category peers are the technical product characteristics highlighted in the figure below.
Firstly, category leaders exhibit substantially higher workflow automation rates, 75% to 90% in some cases, than their peers. This has allowed category leaders to scale and show value faster with fewer resources and capital, meaning they are able to sustainably undercut pricing. That’s one reason why some category leaders are growing over 5x year-on-year, a remarkable rate considering that top quartile growth for healthcare SaaS companies over the last decade has averaged 100% — 150%. A handful of category leaders have actually crossed the legendary $100M ARR threshold in just 3 years, a notable feat for any technology company, let alone a healthcare technology company.
Growth rates of this magnitude are a major driver of valuations. And category leading companies in each of these categories command upwards of a 5x valuation premium relative to their category peers.
Secondly, each of these companies has invested in core product and business attributes that tightly match enterprise buyer preferences. In our enterprise buyer survey, ROI, integration timelines and operational costs were the most important factors considered by buyers evaluating new AI solutions. These preferences are highlighted in the chart below.
The best healthcare AI companies we are seeing today can both execute a sale and integrate their product into core systems of record within six months or less. We have observed that breakout growth is enabled as much by distribution and implementation innovation as it is by product innovation. Our portfolio company, SmarterDx, for example, chose not to integrate directly with medical record systems but instead have their product accessed via web browser, shaving months off implementation. Counter to its scaled incumbent competitors, it also deliberately relied on a contingency-based business model, ensuring the product is essentially “free” for its customers. These are key factors that have allowed the company to contract over $100M in ARR in 3 years. Similarly, Abridge Health announced a privileged co-development relationship with Epic to streamline both integration and go-to-market efforts in 2023. Growth promptly 10x’d the next year. OpenEvidence’s multi-year partnership with the Journal of the American Medical Association (JAMA), the pre-eminent peer-reviewed medical journal, has helped it attract over 50,000 U.S. clinicians a month to its platform, an unmatched pace of growth.
The best healthcare AI companies we have seen also tend to exhibit unimpeachably attributable ROI out of the gate. Category leading AI companies today are exhibiting 5x or greater ROI and some companies can point to this ROI in one to two months after deployment. This is crucial because it means that enterprise buyers can purchase, integrate and see a financial yield on a partnership in about half of a typical budgeting cycle. This affords them the opportunity to reinvest savings before the next cycle, catalyzing the growth of these relationships.
Importantly, leading AI companies do not wait for ROI to manifest. Instead, once the growth flywheel starts, they deliberately manufacture ROI by investing heavily in customer success, low friction user experiences and tight product audits / human-in-the-loop. This ensures consistent product performance and that value attribution can scale over time.
Another important characteristic to discuss is team. In a startup journey, chaos is a constant. Success is ultimately moored by leaders who rely on lived intuition to make millions of correct decisions along the journey. This has been true across technology cycles. And in our view, the best healthcare AI leadership teams possess technical AI depth that is grounded in clinical experience. Across the cohort of companies above, most are led by founders with lived healthcare experience who relate to their end-user and the problem in a deeply personal way. Many combine this experience with inherent technical knowledge or find a complement to it within the co-founding team, enabling deep technical and clinical fluency. But that doesn’t necessarily mean the best founders need to be repeat founders. In fact, amongst category leaders above, most founding CEOs happen to be first-time founders.
Choppy Seas Ahead: Threats and Opportunities for Leading Companies
As category leading healthcare AI companies command soaring valuations, the pressure to grow beyond their initial addressable market is increasing. Not all AI companies are able to outgrow this hype. The ubiquity and accessibility of AI mean that competition and price compression are persistent threats. But, what the category leaders in Figure 7 have in common is that they’ve deliberately positioned their products to take advantage of the target market’s broad surface area and expand ROI. This typically includes horizontal expansion into new categorical areas, a privilege unlocked by trust in the entry wedge and cemented in follow-on contracting structures. The race that all of these companies are now currently running is one to establish new product beachheads sooner than existing ones become commodified.
Perhaps the most frequently scrutinized race for horizontal expansion today is in practice management solutions, where multiple categories of products are vying to become the AI assistant for clinical teams and their administrative staff. AI physician assistant scribes (i.e., Abridge, Ambience, Suki) are actively adding billing, coding and care management features to point to a larger financial ROI story. Meanwhile, leading AI billing and coding companies (i.e., Smarter Technologies, Commure, Codametrix) have bought or are considering launching their own AI physician assistant scribing services to get closer to clinical workflows. There are a handful of other emerging startups also entering this race. For example, given physicians often refer to clinical guidelines during a clinical encounter, it isn’t surprising that OpenEvidence is extending its physician assistant services from clinical decision support to transcription via the expected launch of “Visits”. With so many of these companies vying to become a wholistic clinician AI assistant, we expect a landgrab to solve other major clinical workflow pain points. Some areas that we think will be added to roadmaps in the next year are medication + pharmacy management, patient engagement + portal activities, discharge + care planning and utilization + care management activities.
Incumbent threats also abound for emerging startup leaders. OpenEvidence competitor, Doximity, has already announced its own scribe. Meanwhile, Epic (in partnership with Microsoft) recently announced that it will launch its own AI physician assistant with a suite of scribing, care management, coding and patient engagement features. Epic also suggested that its new AI features would be priced distinctly and incrementally, suggesting a higher level of value and performance. Outsourced revenue cycle management providers like R1 RCM and Waystar (who recently acquired AI clinical documentation improvement company, Iodine, for $1.25 billion) are undoubtedly also evaluating their positioning in this market. These announcements are further accelerating the basis of competition away from any single feature and towards longitudinality.
This dynamic isn’t unique to the practice management market. Clinical imaging and decision support technology incumbents are, for example, fiercely competing with emerging AI startups. RadNet is a good example, having launched its AI-powered portfolio, DeepHealth, in 2023. It has followed that up with a number of imaging AI acquisitions, including iCad, and See-Mode earlier this year. Phillips, also a major player in the medical imaging space, announced plans to invest $150 million into a manufacturing and research and development facility for imaging AI products.
Meanwhile, Evolent, one of the most prominent publicly traded care management organizations, recently announced its intention to become an “AI-first” organization, targeting auto-approving 80% of authorization volume. This accompanies their 2024 acquisition of Machinify’s AI-powered utilization management product. Like many other incumbents, Evolent is signaling that it intends to evolve its technology stack as leading AI startups like Innovaccer and Cohere Health (a Flare Capital portfolio company) reinvent the care management market.
Flush with cash, technological advantages and brand recognition, many emerging startup category leaders are primed to continue scaling with new products. How quickly and effectively they scale clinician and patient stickiness while growing pricing power through extended ROI, will dictate the pace of scale.
Conclusion
AI in healthcare has moved from experimentation to execution. Adoption is broadening, budgets are expanding, and value is concentrating in companies that pair high automation with fast, verifiable ROI. The past year delivered three reinforcing signals:
- Sustained enterprise demand — with many hospitals leading adoption
- Breakaway company performance — some crossing $100M ARR within three years and posting 5x+ year-on-year growth
- Credible liquidity — multiple multi-billion-dollar outcomes validating durable value creation.
At the same time, boundaries between categories are blurring as leading companies expand horizontally, especially across administrative operations where AI’s near-term impact remains most tangible.
Over the next 12–24 months, we expect continued budget expansion, accelerated consolidation, and a clearer separation of winners. Administrative workflows will keep compounding value as products capture and structure “lost” operational data, while the most promising clinical AI will center on decision support built on newly structured clinical and literature corpora — deployed adjacent to live workflows to minimize friction and maximize clinician trust. Distribution will be the decisive moat: privileged partnerships, implementation speed, and pricing models that de-risk adoption will determine who scales.
Strategic Imperatives for AI Buyers and Builders
For providers, payers, and life sciences buyers
- Buy for time-to-value. Invest in internal capabilities that can speed up partnership timelines and facilitate contracting-to-go-live cycles of ≤6 months.
- Establish measurable outcomes. Create infrastructure to enable transparent and traceable ROI attribution between internal efforts and partners.
- Build longitudinally reliable automations. Prioritize partnerships with auditable quality controls and human-in-the-loop guardrails.
- Invest in data exhaust. Favor products that capture, structure, and return operational and clinical exhaust to upstream systems, in order to compound value across adjacent workflows.
- Plan for expansion. Choose platforms with credible roadmaps beyond the entry wedge and align expansion incentives through innovative contracting.
For founders and operators
- Win on distribution and implementation. Treat integration time, change-management, and pricing model design (e.g., contingency-based but only where appropriate) as core product features.
- Instrument ROI from day one. Bake in telemetry, unit-level economics, customer success and clear before/after baselines to drive consistent product usage and expansion.
- Be model-agnostic, data-centric. Build pipelines that normalize heterogeneous data, support rapid model swaps, and enforce provenance, security, and auditability.
- Expand horizontally with intent. Sequence adjacencies that reuse the same data and decisions; resist one-off features that increase integration debt.
- Operationalize trust. Maintain clinical safety reviews, tight QA (including human-in-the-loop), and transparent error reporting.
About Flare Capital Partners
Flare Capital Partners is the leading healthcare technology venture capital firm advancing innovation-driven companies to improve positive health outcomes, broaden care access, and lower healthcare costs. We partner with exceptional founders solving healthcare’s hardest challenges, supporting each with our deep sector expertise, unparalleled industry resources, and proven access to commercial opportunities. We are investors in several leading healthcare AI companies, including SmarterDx, Cohere Health, Suki AI, Layer Health and Axuall Health among others. Our team of established investors and senior operating executives has invested in 70+ companies and has nearly $1 billion in assets under management. To learn more about Flare Capital Partners, please visit flarecapital.com or follow us on LinkedIn, YouTube, or Spotify.
