Tag Archives: artificial-intelligence

That was certainly unexpected. The January payroll report was effectively twice as robust for jobs added to the U.S. economy than was forecasted just several days ago. This was powered by more than 82k and 42k new healthcare and social assistance jobs, respectively, accounting for virtually all of the 130k increase. This strength masks the overall weakness in employment across the other sectors of the economy. It also highlights one of the paradoxes of healthcare technology, which is how quickly automation and AI will lower administrative and clinical costs by reducing labor needs.

Change In Nonfarm Payrolls

Source: U.S. Department of Labor

It is estimated that there are 17 million (and as high as 22 million) healthcare workers in the U.S., according to the National Center for Health Workforce Analysis, accounting for more than 14% of all employees. Foreign born healthcare workers tend to be clustered at both the high- and low-ends of the skill and income distribution, with the lower skilled workers in the home health segment particularly vulnerable to the current Administration’s deportation program.

Foreign workers now account for approximately 15% of the U.S. population but are 39% of all home health aides, according to IPUMS, which tracks census data. In addition to the 230k deportees in 2025, there were another 270k people arrested at the border and another 40k people who “self-deported” in 2025, placing extraordinary pressure on all labor-intensive sectors of the economy.

Estimated Number of Deportees

Source: New York Times

The Centers for Medicare and Medicaid Services recently reported that the national healthcare spending reached $5.3 trillion in 2024, which is an increase of 7.2% over 2023 and is now comfortably above 18% of GDP. On a per capita basis, this is nearly $15,500. There were over 214 million enrollees in private health insurance in 2024, notably seven million more people due to the extension of enhanced Affordable Care Act tax credits. Without these credits, the average ACA enrollee will see monthly health insurance premiums spike from $888 to $1,904, according to KFF.

Against this backdrop there has been a profound shift in the economic composition of the U.S. population with the Middle Class being a smaller percent of the overall population as the affluent class has grown. Across all strata, household income increased significantly over the past half century, although the median income for the lowest income households increased only 55% as compared to 78% for the highest income household, according to the Pew Research Center. In 2022, the ratio of highest to lowest median income was 7.3x (it was 6.3x in 1970). While there may be proportionally fewer poor households, the ability to pay for healthcare has been severely impaired, putting significant strain on public resources and safety nets.

Distribution of U.S. Population

Source: American Enterprise Institute

The importance and magnitude of these social safety net programs is highlighted by an analysis by the American Enterprise Institute that concluded that a family of five with no earnings receives benefits totaling $55k annually. Furthermore, the powerful “leveling effect” of these programs is made more apparent when looking at a comparison of that same family earning $20k versus $80k annually. The higher income family effectively takes home only $11k more than the lower income family in after tax earnings given the substantial government subsidies. Interestingly, Goodwill Industries had its best year ever in 2025 with revenues in excess of $7 billion across its 3,400 stores.

These benefits may contribute to some of the hostility directed at these programs today. For example, Medicaid, which accounts for approximately 19% of all healthcare spending ($919 billion in 2024), is expected to decrease by over $900 billion over the next decade given announced cuts to the federal budget. In so doing, the number of uninsured Americans will increase by more than 7.5 million people.

Furthermore, more than 41.7 million people participated in the Supplemental Nutrition Assistance Program (SNAP) in 2024, according to USDA data. In aggregate, the SNAP outlays totaled $99.8 billion (~$187 per person) and while a federal program, the benefits are administered at the state level which has now introduced a chaotic patchwork system of guidelines and criteria. In certain states, these benefits have become much harder to access.

Hopefully, all these investments in public health are paying off. According to the National Center for Health Statistics, life expectancy in the U.S. hit a high-water mark in 2024 as the opioid crisis started to recede somewhat, dropping by 26%. Those born in 2024 could expect a life expectancy of 81.4 years and 76.5 years for women and men, respectively. The mortality rate in 2024 was 722.1 deaths per 100k.

While 1.2 million people died from cancer from 1990 to 2023 according to American Cancer Society data, the survival rates have improved notably with 70% of cancer patients living for at least five additional years after initial diagnosis. This underscores the criticality of continued robust investment in scientific discovery, and the lunacy of current officials’ attacks on research. Gavi, the Vaccine Alliance, estimated that 19.8 million lives were saved in the first year of the Covid vaccines being available.

In addition to scientific R&D investment, the pace of investment in new healthcare technologies improved in 2025 to $14.2 billion across, an increase of 35% over the $10.5 billion invested in 2024, according to Rock Health. The $4.2 billion invested just in 4Q25, the largest quarterly results for the past ten quarters, suggests a further strengthening of the sector into 2026.

While there were slightly fewer companies that raised capital in 2025 over the prior five years, the average deal size was significantly larger pointing to the impact of large AI financings. One word of caution: there are more than 600 companies which raised capital in 2021 – 2022 that have yet to raise another round. It is unclear what is to become of those companies as many undoubtedly are sub-scale.

Healthcare Technology Investment Activity

Source: Rock Health

There was a marked shift to AI-centric companies by investors in 2025, which now account for more than half of all healthcare technology financings. These companies are focused on solutions to automate many labor-intensive tasks in healthcare. Additionally, healthcare was 22% of all AI financings as compared to just 13% in 2024, pointing to an increased level of investor awareness and acceptance of the potential AI has in healthcare. This also likely reflects companies recasting themselves as AI-forward or AI-native now.

Source: Bessemer Venture Partners

The knock-on effect of AI euphoria has been profound on valuations and round sizes across all stages in the healthcare technology sector. One trend that is clearly evident is the significant “pull forward” of capital to the earliest rounds of financing. Seed investments in healthcare technology AI companies were on average $15.3 million in size at $41.6 million pre-money for an average post-money valuation of nearly $57 million. The pre-money valuation of the following Series A was only a 1.7x step-up, which compares unfavorably to the 3.4x step-up from the Series A to Series B rounds. Even more surprising was the absence of a step-up from the Series B to Series C rounds, perhaps suggesting that many Series B companies struggled to show meaningful commercial progress. 

Healthcare Technology AI Investment Activity (through 1H25)

Source: Proprietary Flare Capital analysis based on Rock Health and Pitchbook databases

The promise of healthcare AI is to materially lessen the reliance on labor in clinical and administrative workflows. In spite of essentially flat Medicare rate increases announced for 2027, there are several promising federal initiatives to facilitate greater efficiencies such as the ACCESS Model (AI-supported care models for chronic disease) or the WISeR Model (AI-enhanced prior authorization) or the MAHA ELEVATE Model (evidence-based functional lifestyle medicine interventions). Quite clearly, the current Administration is pushing to remove many of the AI guardrails in place to encourage business model transformation.

One interesting development has been the extraordinary level of chatbot engagement for health-related matters. OpenAI recently rolled out a health tab in ChatGPT that now has 40 million daily visitors. Patients are uploading medical records and test results, desperately looking for guidance and insights while circumventing already capacity constrained providers. This phenomenon underscores the opacity and frustration of engaging with the healthcare system today that operates like a gated community as patients scurry into corners of the internet looking for answers. It also highlights the need for greater “data liquidity” so that fringe commentators do not receive media attention in the absence of authoritative clinical voices.

Interestingly, it is not just patients looking for online assistance. Physicians are now readily embracing AI tools at the point-of-care. Given the pressure on healthcare labor, such advances, if reliable, will provide significant leverage for the provider workforce.

Most Popular AI Tools for Physicians

Source: 2025 Physicians AI Report

Notwithstanding the precipitous drop in consumer confidence this past month from 94.2 to 84.5, the lowest monthly reading in a dozen years and lower than at the depths of the Covid pandemic, public market investors seem to be returning to the healthcare sector. Ironically, general consumer sentiment appears to have soured on the state of the jobs market, although jobs remain relatively plentiful in the healthcare sector even with rapid AI adoption. According to national net buy/sell trading data, an analysis by Morgan Stanley concluded that of the eleven tracked sectors, healthcare experienced the greatest month-over-month improvement in overall investor sentiment to start the year.

Monthly Sector Rotation

Source: E*Trade, Morgan Stanley

We hope you will join us for our next quarterly Expert Roundtable Series on March 3, 2026 at 12:00pm ET (register here). To stay connected, subscribe here for the latest updates, news, and insights from Flare Capital Partners.

Investments in healthcare technology should work. There can be a legitimate debate about the wisdom of the magnitude of overall AI investment activity, now estimated to be $320 billion just in 2025, but undoubtedly most will see significant healthcare benefits from this infrastructure buildout.  But there may also be merit to the argument that certain knock-on effects of AI will undermine for some improvements in the quality of life. At a minimum, it is not entirely clear that the agentic arms race unfolding between payors and providers will entirely help patients versus improving the economics of the combatants.

Not surprisingly, the geographic diversity of the U.S. between urban / suburban / rural directly influences one’s life expectancy. The Economic Innovation Group has developed a Distressed Communities Index which is a composite of factors including poverty levels, education, housing, median income, and employment status. Just over half of Americans live in suburban zip codes, and nearly two-thirds of suburbanites reside in what are deemed Prosperous and Comfortable areas. Those fortunate residents have life expectancies nearly eight years longer than those living in Distressed areas. Just over 15% of Americans live in Distressed zip codes.

Life Expectancy by Distress Score

Source: Economic Innovation Group

Bank of America estimates that those earning more than $250k represent approximately 10% of all Americans and now account for over 49% of all spending, a significant step-up from 35% just 30 years ago. This will be enormously useful for those more fortunate given healthcare cost inflation has spiked according to recent Consumer Price Inflation data. Annual hospital costs and prescription drug costs are expected to increase 6.8% and 7.0% in 2025, respectively, according to Peterson-KFF forecasts. KFF goes on to project that the cost of an annual family health insurance plan is now nearly $27k, up 6% from 2024 levels.

Next year is likely to be even worse. Large employers are expecting overall health costs to increase 9% in 2026, in large measure due to a 12% jump in drug spending (oncology, GLP-1 being principal contributors). A survey from the Business Group on Health expects employers to meaningfully cut costs (versus passing increases on to employees), perhaps under the aircover of expected AI efficiencies.

Therefore, many Americans are now highly reliant on an array of important social services with nearly 24% on Medicaid or the Children’s Health Insurance Program (CHIP) – that is 85 million people. It is estimated that up to 16 million will lose health insurance coverage by 2034 due to the One Big Beautiful Bill Act. Should the government shutdown continue into November, an estimated 42 million low-income people will not receive food stamp benefits from the Supplemental Nutrition Assistance Program. As many of these programs are now under siege, the expectation (i.e., hope) is that investments in healthcare technology and AI will be able to improve care services for those most disenfranchised, living in Distressed communities to extend or augment the existing healthcare infrastructure.

Percentage of Country Receiving Social Benefits

Source: Census Bureau, USA Facts (2022)

Notwithstanding the significant costs incurred to provide care to Americans, estimated to be over $12k per capita, life expectancies are middling and on parity with many developing countries. This is not something to be particularly proud of in light of how much is spent on healthcare and technology – and points to the profound opportunities to re-architect the healthcare system.

Per Capita Healthcare Expenditures vs. Life Expectancy

Source: Silicon Valley Bank

Perversely, the dramatic investment in AI infrastructure may have significant negative health implications for some due to increased social isolation and loss of many jobs. A recent Financial Times analysis concluded that specifically young men and the unemployed exhibited a significant – and recent – spike in “deaths of despair.”

Percentage of Men Reported Feeling Isolated

Source: Financial Times, BBC Loneliness Experiment

Those who reported being isolated were nearly 3x as likely to die from substance abuse or suicide. The heightened incidences of online porn and gambling addiction have been widely reported. The prospect of widespread job loss due to AI will be devastating to those who are unable to adjust, especially in light of compromised social safety nets. The CEO of Heineken recently argued that his alcoholic products have health benefits given they are “social lubricants” which help people feel less isolated. This has been reflected in the company’s marketing campaigns for generations.

In spite of some of these concerns, the healthcare sector largely has embraced AI and identified bespoke solutions to transform administrative and clinical workflows. Today there is an agentic arms race between providers and payors pointing AI agents at one another to optimize reimbursement payments. A survey by Menlo Ventures concluded that 22% of all healthcare organizations have secured commercial AI licenses (versus 9% overall). Importantly, though, 27% of providers have as opposed to only 14% of payors, highlighting the current imbalance in automation capabilities.

An analysis by Bain highlights the continued priorities around revenue cycle management (e.g., documentation, coding, billing) and clinical workflow optimization, code for improving cash flow and reducing operating costs. Investments that do not directly have immediate ROI (~3-5x in first year) tend to fall lower down the stack of needs. Notably, the urgent requirements at the start of the pandemic to have providers become fully on-demand, always on, and virtual (e.g., telehealth and patient engagement) are meaningfully less critical now.

Survey of Current Provider Budget Priorities

Source: Bain

A similar Bain survey of payors shows a distributed set of priorities focused on product design (network and plan design) and member coordination. Many of the other priorities impact issues involving payment and cash flow and lend themselves nicely to being automated. In the face of increased medical loss ratios, a spike in utilization rates, lower reimbursement, and rapidly shifting member eligibility payors are aggressively considering AI solutions. Better determining provider quality will improve network design such that members will access more appropriate (i.e., likely less expensive) care at the right time and setting, improving member experience and reducing churn. Ironically, member-facing activities tend to be the first use cases to be automated (e.g., call center, member engagement, benefits verification).

Survey of Current Payor Budget Priorities

Source: Bain

After a challenging period post pandemic, this year many of the healthcare public stocks have performed reasonably well and for some sectors such as hospitals and large service providers, they have significantly outperformed year-to-date some of the broader market indices. Of note, the Medicaid sector has struggled as have smaller providers (sub-scale), risk-bearing entities, and tech-enabled virtual care companies, according to an analysis by J.P. Morgan.

Sector Stock Performance (2025 YTD)

Source: J.P. Morgan, Bloomberg

At the end of 3Q25 the Healthcare Technology Public Company Index (46 companies) tracked by MTS Health Partners was trading at an average of 3.0x and 2.8x 2025 and 2026 revenues, respectively, and 17.0x and 15.2x 2025 and 2026 EBITDA, respectively. The Index was forecasting an average 7.1% revenue growth for 2026, a marked slowdown from the 9.3% increase over last year, reflecting many of the industry headwinds. The most highly valued sub-sector was the Pharma-Tech cohort, which is valued at 5.1x and 16.6x 2026 revenues and EBITDA, respectively. Value-Based & Tech-Enabled Primary Care and Tech-Enabled Payers were the two lowest valued sub-sectors at the end of the quarter, likely reflecting the capital intensity of risk-bearing business models.

Investor excitement in healthcare AI is reflected in digital health investment activity, notwithstanding challenges to reconcile the more modest valuation multiples for publicly traded companies with those of lofty private market multiples. According to Rock Health, year-to-date $9.9 billion has been invested in 351 companies. For 3Q25, $3.5 billion was invested in 107 companies putting the sector on pace to have ~$14 billion invested in around 450 companies, which would signify an encouraging recovery closer to 2022 levels.

There continues to be evidence of further capital concentration around emerging winners in the digital health sector. Average investment size increased significantly year-to-date to $28.1 million from $20.4 million last year, driven by the dramatic increase in round size for AI companies. The top three focus areas in 2025 are (i) clinical workflow (ii) non-clinical workflow and (iii) care coordination, according to Rock Health. Silicon Valley Bank estimated that 38% of healthcare technology investments this year were in “mega rounds” of over $100 million, up from 28% in 2024.

Digital Health Investment Activity

Source: Rock Health

Expect further consolidation among the great unwashed masses of sub-scale private healthcare technology start-ups, which will ultimately be healthy for the sector. Going forward fewer companies will likely be started against a backdrop of more capital being invested in the sector. In addition to the resilience in the pharma-tech sub-sector, there is increasing investor chatter around several emerging themes such as the development of proprietary healthcare data sets (versus building on top of general purpose LLMs) and “hardware delivered AI.”  Never a dull moment…

It is hard to stomach the fact that there are an estimated 60 – 70 million Americans who suffer from gastrointestinal (“GI”) conditions, costing the U.S. healthcare system approximately $140 billion annually, according to the National Institute of Diabetes and Digestive and Kidney Diseases. BCC Research tallied the costs of just the therapeutics and associated diagnostic procedures were $86 billion last year, increasing to more than $123 billion by 2028. There are now more than 100 million GI ambulatory care visits each year and GI conditions are thought to be one of the four most expensive diseases.

A recent WTW (formerly Willis Towers Watson) study concluded that 72% of Americans experience GI symptoms multiple times each month, yet only 25% are properly diagnosed. More than 40% of the U.S. population state that their daily lives are “disrupted” by GI issues, contributing to the second most prevalent short-term disability claims in 2020.

GI conditions are often misunderstood and include a complicated basket of conditions including irritable bowel syndrome (“IBS”), inflammatory bowel disease (“IBD” – includes Crohn’s disease and ulcerative colitis), gastroesophageal reflux disease (“GERD”), celiac disease, and diverticulitis, among many other associated issues (gallstones, various cancers, etc). There are over 700 GI-related ICD-10 reimbursement codes. While complicated, this is not surprising given the anatomical terrain covered.

Given the heterogeneity of GI conditions, diagnosis and subsequent treatment guidelines are complex and often characterized by numerous false starts. Patients are often floating between primary care providers, who are often at a loss as to underlying issues, and specialists. Exacerbating the patient’s journey are issues of access, wait times to schedule, and required follow-up care. These delays often lead to increased ER admissions and more expensive interventions.

It was into this vortex that Oshi Health was founded. Flare Capital co-led the Series A financing in late 2021 and the company subsequentially raised a larger Series B in March 2023 to scale the clinical model.

In part as a function of the pandemic when the healthcare system had to become virtual/on-demand/predictive/always-on overnight, there was a dramatic influx of capital in 2021 of nearly $30 billion invested in healthcare technology companies, much of which went to launch virtual specialty care models across many diseases. Notably, there are now several dozen behavioral health companies, but other specific diseases such as kidney care, cardiology, and oncology are now in an “emerging winners” phase of the market as many of those venture-backed companies are now scaling to be substantial businesses.  

As Oshi Health scales, there are some interesting observations and learnings that have emerged that might guide other specialty care models.

Size of Market Matters: The GI market is unambiguously large and woefully underserved – and is very important to payors (see below) who are managing to very strict medical loss ratios and benefit plan designs, even more so in today’s economic climate. Solutions addressing smaller patient populations will struggle to get payors’ attention. The size of market calculus needs to also reflect the ease of implementation. One must not underestimate the cost/benefit analysis that payors will do prior to implementation.

Team Matters More: Congrats to the CEO, Sam Holliday, for building a terrific leadership team. Sam recognized the need, actually requirement, to have deep clinical expertise early in the team building process. He also leaned hard into building a depth of analytical and actuarial talent early to help “tell the story.” These capabilities have been complemented by excellent financial leadership which facilitates insights around optimizing the care model (see below).

Need for Outcomes Data: The team smartly ran a clinical trial early in the journey which underscores the need to show outcomes data. These data supported the business case that the company both avoided unnecessary (and expensive) ER visits and diagnostics procedures while lowering costs. In order to get credit for that impact, the company needed to have validated data that attributed the impact to Oshi Health’s clinical model. The actuaries at payors needed to scrub those data as well. There is a tension here as clinical trials can be expensive, contributing to the capital intensity concerns.

Novelty of Value-Based Care Models: While there is much discussion of these VBC models, be forewarned that in many cases payors have not yet deployed the appropriate claims infrastructure to even process reimbursements, particularly when existing codes do not neatly map to the milestone-based approaches of some of these care models. This is a big deal and could easily add several quarters to successfully scaling the company. For example, principal care management codes tend to be billed monthly, yet VBC models may have fewer submissions tied to outcomes. Can these models really get credit (i.e., paid) for avoided ER visits? Not to be overlooked are other secondary points of friction that involve the fraud, waste, abuse and payment integrity teams within payors who also need to bless these arrangements.

Balancing Supply and Demand: The business model demands excellent execution across dozens of elements, none more important than scaling the clinical supply with patient demand. This is perhaps the most challenging dimension as the company will need to build supply in advance of demand. Setting measurable interim operational and financial milestones around provider utilization, patient satisfaction, enrollment, conversion, time to book, number of interactions per patient, etc will illuminate the pressure points of the care model. For instance, excess capacity can be hard to see but is brutally expensive. Even slight delays in appointment booking directly hit revenues.  

Tech-Forward Approach: Every element of the care model affords opportunities to optimize with technology. The Oshi Health team has really leaned into deconstructing each step of the care journey to identify areas that technology can improve upon, particularly around care coordination. In fact, the company recently hired a Chief Technology Officer to further drive its commitment to be tech-forward.

Capital Intensity: Scaling a national specialty care model is expensive. While it may be possible to launch in a specific geography, large employers will want to see national capabilities. While a precise threshold is hard to determine, a reasonable working assumption should be that this will be $100 million endeavor to build a robust national offering. Therefore, the role of strategic investors cannot be understated and should be embraced early. Oshi Health has an extraordinary relationship with the team at CVS Health, an important partner and investor, which has been a significant resource for the company. Additionally, the company has ensured depth of follow-on investment capacity within the existing investor syndicate.  

Virtuous Cycle: Good things happen when the model works, creating this powerful “flywheel effect.”

While not an exhaustive list, these are a handful of dimensions to scaling the company which the team has been particularly effective at managing. It has been exciting to see the accolades that the company has so deservedly won. Last month, Inc. Magazine cited Oshi Health as “One of the Best Workplaces,” while in May, Modern Healthcare called out Oshi Health as “One of the Best Places to Work in Healthcare” as well as its “Top Innovator.” In February, Fierce Healthcare listed Oshi Health as one of its “Fierce 15” Companies for 2024.

These accomplishments are easy to swallow…