Sponsored by: Limina

Everyone is racing to put clinical data into AI but fewer are sweating the PHI buried in it. It’s rarely the obvious PHI that creates the most risk; clinical context, implied identity, and sensitive conditions are much harder to detect.

Limina’s benchmark report found general-purpose cloud tools miss 13.8% to 46.5% of PHI. An independent audit measured Limina at 99.5%+ accuracy. No surprise, then, that companies like Doximity, Spring Health, and Providence Health choose Limina.

The benchmark report is pretty eye-opening. Check it out here.

If you're interested in sponsoring the newsletter, let us know!

AI

How many $100 billion AI healthcare admin platforms will there be?

I thought one of the more interesting financing rounds of the last few weeks was Commure’s $70 million funding round at a $7 billion valuation. Two things in particular stood out to me:

  • GC’s Hemant Taneja posted on LI that Commure needs to become a $100 billion company to fulfill its potential as it attacks US healthcare’s $1+ trillion in administrative spending

  • Commure’s CEO, Tanay Tandon, appeared on the Sourcery podcast, invoking the Rockefeller-esque approach of building a dominant platform via acquisition while, in his words, eviscerating point solutions with a passion. In the same conversation, Tandon noted that Commure often wins contracts with health systems as a point solution.

Tandon and Taneja make it pretty clear that the ambition here is for Commure to become the winning platform for healthcare AI, leveraging its market position with systems like HCA and Tenet to roll up various point solutions on its way to becoming a $100 billion business. In a market where tech AI companies like Anthropic, SpaceX, and OpenAI are commanding valuations in the $1 trillion ballpark, you can see the conceptual appeal of winning the race to apply AI to an industry that represents ~20% of US GDP.

I think it’s an interesting thought exercise to ponder the likelihood of there being a $100 billion AI infrastructure company for the US healthcare system in the future. Commure seems like one of a handful of big bets on building this platform, and perhaps the most ambitious in its M&A strategy, but far from the only effort here.

Tandon notes in the Sourcery interview that this market will inevitably consolidate over the years ahead, creating carnage in the point solution market, as well-funded startups that don’t achieve scale will see contracted revenue evaporate as customers migrate to platform offerings over time. This does seem like a natural evolution in the market, and a safe bet on how many healthcare AI startups will fail. While it seems clear that most point solutions on the market today won’t make it, it seems harder to predict which platform will emerge as the winner in the market.

Given that, I found it fascinating how Tandon described Commure’s deployment at Summa Health (the Ohio-based health system that GC now owns) in the interview — essentially, Commure has a group of consultants Forward Deployed Engineers on the ground at Summa identifying efficiency opportunities, with Summa serving as a test case to show that AI can save a failing health system. It’s the clearest articulation of Commure’s value proposition I've heard to date. It also sounds a lot like the value prop consultants have been pitching to health systems for ages (apparently with little success).

The bull case for that model probably sounds a lot like Palantir-for-healthcare. On the other hand, Health Catalyst is a pretty direct comp in the public markets, and its $100 million market cap provides a very real bear case scenario for these companies. Not to mention the challenges that have befallen previous attempts at solving this problem (i.e. Olive AI).

Personally, it doesn’t seem crazy to me that a decade or two from now, we’ll look back on this period as one in which a new $100 billion healthcare business takes shape, similar to how UHG began to really take off in the late 2000s. I also don’t think that company will remotely resemble the Palantir-for-healthcare model. It seems inevitable to me that a $100+ billion company would need to own more of the care delivery / financing dollar to achieve that type of scale, similar to UHG / CVS / HCA today. Either way, it’ll certainly be interesting to watch how Taneja and Tandon channel their inner Rockefellers in an effort to acquire their way to that type of outcome.

GLP-1s

GLP-1s and their impact on obesity rates

Epic Research released a fascinating dataset, examining BMI over the past five years. Given the cultural impact of GLP-1s in 2025, plus the fact that Eli Lilly and Novo Nordisk generated around $36 billion in revenue in 2025 from GLP-1s, I figured we might see a noticeable shift in the distribution of the adult population towards healthy weights from overweight / obese. I was wrong — here’s the full adult population in Epic between 2021 and 2026:

Source: Epic Research

Individuals with healthy BMIs have increased by 0.5% over the past five years, with overweight and obese BMIs declining by 0.5%, from 73% of the population to 72.5% of the population, which is hardly noticeable on the chart above. If you look at the Epic data, there is a more pronounced shift in the population taking GLP-1s, which saw a 2% increase in healthy BMIs between 2024 and 2026.

Still, if the general theory is that GLP-1s will solve obesity, and 12% of the population reported taking a GLP-1 in 2025, that will need to show up in a chart like the above at some point, right?

We spent some time in Slack this week talking through how it can be true that: 1. that chart shows no sign of movement, 2. GLP-1s generated $36 billion in revenue in 2025, and 3. polling shows 12% of the population took them.

CLINICAL AI

Early Utah AI pilot data provides a glimpse into rollout

The State of Utah released a report on the early results from Doctronic’s AI pilot. Thus far, AI has recommended approval for 72% of prescriptions, and 91% of the time, a physician has approved that recommendation without requiring any further information. 28% of the time, the AI escalated to a physician, and in 31% of those cases, the physician found that the AI was being overly cautious in escalating it.

Source: Utah data; HTN analysis

Two things stick out to me here:

  1. For all the concern this pilot has generated about AI replacing physicians, a human still appears to be in loop in every part of this pilot process and those concerns seem overblown. Given all of the oversight here, I’m hard-pressed to imagine this process is not significantly safer than the current state, despite the concerns.

  2. Utah noted the number of patients going through this pilot is still quite low. That issue of patient adoption seems like the more interesting data point to keep an eye on, particularly with the state medical board pushback.

Sponsored by: Abridge

On June 11, Abridge is hosting its first-ever Keynote, live from NYC and streaming globally. Abridge is on a mission to save time, save money, and save lives. Join online to learn about the most significant step toward that goal to date. Register here.

News items on hot topics we’re tracking in HTN

  • Clover Health won a potentially major Stars lawsuit against CMS.

    • Clover’s stock was up 16% on Thursday as a judge ruled in Clover’s favor, throwing out a number of Stars metrics. CMS is appealing, but if it stands, this bumps Clover’s plan from 3.5 Stars to 4 Stars, resulting in ~$120 million in bonus payments in 2027. Andrew Schwab speculated on LI that between this and the Chevron ruling, the outcome here could portend bigger changes ahead for the Stars program.

  • Oura joins the ranks of IPO candidates while hinting at healthcare ambitions.

    • Last week, Oura took a major step forward in the IPO process, confidentially filing its S-1. This week, in made some major product updates, including adding a number of health-related features to its device (blood pressure monitoring, nighttime breathing, GLP-1 tracking) as well as a partnership with Counsel Health to provide in-app access to AI-enabled care delivery. The Counsel Health partnership is particularly interesting here, as it makes a trend with Whoop and now Oura integrating clinician access into the experience. While Oura is not currently participating in CMMI’s ACCESS Model, Abe Sutton indicated that more participants will be added to the model soon, and one would imagine Oura would make for a good candidate.

  • The No Surprises Act continues to generate headlines, with providers pushing back over payer (specifically BCBS Texas) non-payments

    • Last week, Radiology Associates of North Texas issued a press release noting that BCBS TX has paid only 2% of the binding awards from the IDR arbitration process, and that it will spend $51+ million in administrative waste tied to the process. This week, HaloMD issued a press release announcing that a federal court dismissed BCBS Texas’s complaint. The law might not make any sense to either side of this argument, but nonetheless, it is the law. Also this week, CMS made some adjustments to streamline the IDR process.

  • Regional health plans continue to struggle as two plans in the Pacific Northwest shutter

    • Providence Health Plan and PacificSource both announced plans to exit insurance markets. Providence is shutting down its ACA, Medicaid, and employer businesses in 2027 while partnering with a national carrier on Medicare Advantage. The Williamette Week provided a sober assessment of the situation at Providence, highlighting perceived mismanagement by a leadership team more focused on startup partnerships than running a health plan. Regardless of the reasons, the health plan lost $100 million in 2025, leading Providence to shut down. PacificSource, a non-profit regional health plan partly owned by Legacy Health, is exiting the ACA market entirely, including all lines of business in Montana, affecting ~42k of PacificSource’s ~500k members. This move comes after multiple attempts over the last several years to rightsize the business. It’s another sign of the headwinds facing regional health plans in the current environment, although perhaps Q1 2026 earnings calls provide a glimmer of hope that 2026 will be a better year.

Other Headlines

A summary of other key headlines to be aware of

  • WakeMed declined a $5 billion merger offer from UNC Health, opting for its deal with Atrium Health despite UNC’s offer being more than twice as much. WakeMed leadership noted that it is a competitive disadvantage because it charges 10% - 30% less than others in the market, which perhaps suggests what will happen to healthcare costs after a transaction here.

  • CMS sent a letter to Elevance on Friday, noting that Elevance completed the second of four steps it needs to finish by June 30th in order to avoid enrollment sanctions.

  • Amazon Health Services brought in a new leader, Dr. Roy Schoenberg, the co-founder of AmWell. Given how often I’m now seeing advertisements for Amazon virtual care, it seems indicative of the strategy moving forward.

  • The Massachusetts AG is suing UnitedHealthcare over $100 million in alleged upcoding in its D-SNP program between 2014 and 2025. For folks in the duals market, the case is a good read on just how complex these programs are.

  • Open Evidence partnered with Cedars-Sinai to integrate patient data into the OE search experience. Seems like a cool move for OE, particularly as it relates to about an enterprise sale into health systems. Worth keeping an eye on implications for OE’s go-to-market strategy, particularly given this coincides with Doximity talking about how the pharma advertising market has been soft recently.

  • WISeR continues to face pushback in DC, with House and Senate Democrats calling for an end to the program.

  • CVS filed a lawsuit against Tennessee over a new law barring a PBM from also owning a pharmacy in the state.

  • HCA announced an agreement to acquire The College of Health Care Professions, which provides healthcare education to 8,000+ students annually across ten campuses in Texas and online.

  • K Health partnered with UPenn Medicine for an AI-centric virtual care offering. Given all the excitement around AI for care delivery, K Health’s relatively quiet momentum with partnerships at UPenn, MGB, Northwell, and Mayo seems notable.

  • Innovaccer acquired CaduceusHealth to expand its revenue cycle capabilities.

SVP, People at Pearl Health, a primary care physician enablement platform.
$250K - $290K | Hybrid (NYC)

Head of Product Operations at FamilyWell Health, women’s mental health embedded into OB/GYN practices.
$135K - $160K | Remote (East Coast preferred)

Policy Senior Associate at Morgan Health, a division of JPM focused on improving employer-sponsored healthcare.
$115K - $140K | In-office (DC)

Enrollment Operations Lead at Givers, supporting and paying family caregivers.
$150K | In-office (NYC)

VP of Data at 9amHealth, a digital health company focused on cardiometabolic care.
Remote

Contact us to feature roles in our newsletter.

Funding Announcements

Highlighting a few of the most interesting funding rounds from the week

  • Garner Health, a provider of data analytics platforms for employers, raised $100 million at a $2.74 billion valuation.

    Index Ventures led the round. This comes three months after it announced a $118 million Series D at a $1.35 billion valuation. Garner cites 800 customers and $200 million of ARR, which has more than doubled each of the last five years. Per Bloomberg, revenue has apparently also doubled since the $1.35 billion valuation three months ago, which doesn’t quite compute. Either way, the general sense is that Garner has a lot of momentum with employers. Garner’s pitch increasingly sounds like a healthcare navigation play for employers, with the ultimate proof point being helping them drive 12% cost savings. Given the situation with employer costs, it’s not exactly surprising that we’re seeing a lot of activity in this market, and I’ll be keen to keep an eye on if we start to see consolidation in this market in the months ahead as winners start to emerge.

  • Nourish, a virtual dietitian-led clinical model, raised $100 million at a $1.75 billion valuation.

    The round was led by Menlo Ventures, bringing Nourish's total capital raised to $215 million. Menlo noted in its blog post about the round that it thinks Nourish has cracked the patient acquisition problem by leveraging a pharma-style field sales motion with sales reps turning providers into referral engines for Nourish’s care model. Between this and the mental health market, it is interesting to hear more startups picking up on this notion of providers as a key go-to-market strategy.

  • Signos, a D2C continuous glucose monitoring brand, raised $20 million.
    The round was led by Cheyenne Ventures and GV, with participation from DexCom and BCBS Alabama. Dexcom’s involvement seems notable as it manufactures its own D2C CGM, Stelo, which Signos uses. Dexcom is under pressure from an activist investor (Elliott Management) over a lack of execution recently.

  • Kin, an app that helps patients record provider visits, raised $9 million.
    Maveron led the round, with participation from Town Hall Ventures and others. As I’ve noted here before, this model sounds conceptually similar to where Abridge started out. The timing of this round with Shiv Rao’s conversation on 20VC about Abridge’s “five-year desert to product market fit” recently felt serendipitous. We chatted with Kin’s Co-Founder and CEO, Arpan Parikh, about how they’re contemplating the business model for a free patient-facing app, and it very much sounds like the thesis centers around coordinating healthcare benefits for individuals and solving the “last mile” challenge in healthcare. Worth noting that the GoodRx co-founders are also co-founders of Kin, and I’d expect we'll see a similar model deployed here. Certainly an interesting twist on a cool concept that is worth keeping an eye on.

  • Paralign Health raised $3 million to build a network of paramedics providing preventive health services in rural parts of the US.
    Flyover Capital led the round. Paralign helps fire departments and EMS agencies to implement Mobile Integrated Health and Community Paramedicine (MIH-CP) programs operationally. Our understanding is that in rural environments, much of care delivery already falls to fire departments/EMS, and something like this makes sense if the financing can be sorted out.

What I’m Reading

A smattering of other interesting pieces from around the web

  • Last week, I spent some time checking in on Ezra, the D2C MRI brand that Function Health acquired a year ago. At the time of acquisition, Function noted that Ezra expected to be in 1,000+ clinics within a few months at a $499 price point. Function’s website this week notes it is available in 189 locations at a $999 price point. Seems like a clear indicator of slower-than-expected growth in that D2C market, given Ezra’s model is based on negotiating volume discounts from outpatient radiology providers like Akumin.

  • The WSJ riled up the online physician community the other weekend by penning a piece suggesting that Nurse Practitioners are the hottest job in healthcare because NPs can provide “doctor-like” services at a lower cost. Like most topics, I understand there is nuance to this conversation, but given the current state of healthcare access and costs in this country,

  • A JAMA Network article arguing that we should treat primary care as a public utility provided some interesting fodder for a Slack conversation about what it means to think of primary care as a public utility. The reaction to the WSJ piece above hints at the challenge of defining what access to primary care means (an AI chatbot? an NP? a MinuteClinic location? a physician?) Part of the challenge at hand is that I think the average American’s answer to that question is likely quite different from how a primary care physician might answer it.

  • Lest you think access to primary care is a uniquely American challenge foisted upon us by local monopolists and vertically integrated conglomerates, the Financial Times reports that the NHS is finalizing a workforce plan that will involve more AI for care delivery in order to help stabilize the workforce. Predictably, this is already being railed as a “massive, dangerous gamble” by the British Medical Association.

Want to comment or share feedback?

Login or Subscribe to participate

If this newsletter was forwarded to you, subscribe here or see more from Health Tech Nerds.

Reply

Avatar

or to participate

Keep Reading