Weekly Health Tech Reads 9/13/26

Today’s newsletter is brought to you by: Champ AI

Champ AI automates what APIs can’t reach. They identify, build, and run multimodal workflows end-to-end — it's a single platform that reliably works across browsers, portals, unstructured documents, and phone calls, covering everything an operations team touches in a given day.

Their custom approach helps organizations close the “last mile” gaps that often block AI adoption. Learn more about how healthcare companies like Pair Team and Float Health use Champ AI to automate operations.

If you're interested in partnering with Health Tech Nerds, let us know!

It’s been interesting to juxtapose the social media reaction to the recent Emanuel, Khosla, et al JAMA article on autonomous AI with ARPA-H’s announcement that it has committed $33.7 million to test autonomous clinical AI for cardiovascular care, via its Agentic AI-EnableD CardioVascular CAre TransfOrmation (ADVOCATE) program. This chart provides a summary of the entities involved:

Yet while my social media feed remains littered this week with various clinicians pillorying the JAMA perspective, seemingly equal parts concerned about the safety and viability of the concept of autonomous clinical AI, there’s been virtually no reaction to ADVOCATE, which is actually working to implement the concept of autonomous clinical AI in partnership with both startups and leading institutions including Stanford, Kaiser, Duke, and the AHA.

I look at that slide, and think to myself, man, ADVOCATE seems like a very cool approach to developing clinical AI. It combines support/funding from the federal government, deployment support from large institutions, and startups building out new agentic AI capabilities. This seems like exactly what we should be cheering for as an industry. Why isn’t anyone talking about this more?

It’s a bit odd to me to see a theoretical conversation about this concept generating such blowback, while a practical version actually deploying it goes bump in the night. Take UpDoc as an example of an awardee. It received $9.2 million to develop its clinical AI for autonomous heart failure management from ARPA-H. UpDoc previously announced its FDA-approved autonomous AI model earlier this summer, noting that institutions like Cleveland Clinic are rolling it out.

Personally, it is a pretty clear indicator to me that the concerns raised in response to the JAMA article are a bit overblown. I had the chance to interview Zeke last Friday, and I asked him about the future state of how people access care in this world of autonomous clinical AI. From where I sit, his answer sounded a lot like the general path of travel that ADVOCATE appears to be taking.

The ARPA-H approach strikes me as very interesting here, and when coupled with the conversation with Zeke, both seem to point to the general path ahead for the development of autonomous clinical AI: healthcare delivery institutions will manage AI agents acting on their behalf to autonomously handle a set of well-suited, well-defined use cases. Couple the technological advances with the fact that leading academic institutions have always viewed primary care as a loss leader necessary only to ensure downstream volume, and I’d argue there’s also a pretty clear financial rationale (and probably clinical, too) for leading academic institutions to head in this direction.

I suspect that one of the primary drivers of the different reactions to the JAMA piece and the ADVOCATE announcement is that JAMA invites a question that naturally challenges many clinicians: what is the role of the clinician in an autonomous AI world? I think ADVOCATE elicits less of a reaction because the clinician's role still seems pretty clear to me here. What one person may call autonomous clinical AI, another person may call the next generation of remote monitoring (or digital therapeutic if you’re so inclined). The latter seems much less offensive to the clinical community, understandably.

This shift strikes me as a huge opportunity for enterprising clinicians thinking about what the practice of the future looks like, and I’d actually be willing to bet that if clinical AI works, it will be a good thing for clinicians, much the same as the radiology market has played out over the last decade.

Then again, who knows, apparently AI will kill us all in the next ten years, so perhaps we should stop fretting about this and throw our support behind Bryan Johnson in unlocking immortality.

Everyone reading this newsletter is surely aware at this point of the conversations about cost increases facing the employer market. Politico again highlighted this conversation on Friday, discussing rising healthcare premiums as the next “cost crunch” facing workers. I very much get the sense in conversations that these cost increases are finally resulting in an “in case of emergency, break glass” moment for employers, where they are considering making more drastic changes to healthcare coverage.

Those changes appear to be coalescing around a handful of ideas that seem to be steering employees toward higher-quality, lower-cost options, using a variety of approaches. Those handful of approaches seem to be at the top of this list from the recent Pulse of the Purchaser survey by the National Alliance of Healthcare Purchaser Coalitions, highlighting how employers are seeking to manage hospital costs:

a16z’s Julie Yoo captured this general vibe well in a blog post on Thursday, highlighting this as a generational moment in time she refers to as “the great health plan replacement,” with the table below highlighting how employers are rethinking core health plan concepts, with a few example startups in each category.

One of the more interesting components of this slide is that the last era of approaches to digital health innovation for employers — think brands like Omada, Hinge, Sword, etc — is completely missing as a category in this table. Why is that?

I’d suggest that going back to the Pulse of the Purchaser survey provides some indication here. That chart I included above is from a section on how employers are addressing hospital costs. It would seem that is the core issue at hand, and employers are pursuing these strategies in response to rising costs, while hoping that steering care more deliberately toward high-quality, low-cost care delivery settings can help better manage healthcare costs. While many leading digital health companies have matured over the past several years, employers are still left grappling with these cost issues, offering a sobering glimpse into how relevant those solutions are in today's conversation and the potential headwinds ahead for the traditional digital health models.

At the same time, I’d imagine all of this disruption makes for an exciting time if you’re one of the “challengers” on the list above. If you’re looking for the next generation of “rocket ships” in the employer market, I’d expect they reside somewhere either on or adjacent to this list. Garner Health seems to epitomize this, with its valuation now at $2.74 billion following a May 2026 Series E financing. Note that valuation doubled over the course of a few months from its February 2026 Series D valuation of $1.35 billion. It seems like a safe bet that we'll see a handful of companies on that list have very nice exits over the coming years as the market reacts to this broader shift and the new capabilities required to win.

When handicapping what companies will ultimately succeed here, I think it’s worth taking a step back to consider the capabilities health plans will need to win in the employer market in the coming years. As employer appetite for steering care increases (I’m using steering in a very broad sense here), the BUCA (BUCA = Blues, UHC, Cigna, Aetna) incumbents will almost certainly react and begin adopting these plans. Figuring out what they’ll choose to acquire, partner with, or build internally will be key. We’ve already seen this playing out in the dynamic co-pay market, where it strikes me as a bit ironic that perhaps the most successful “challenger” in the market is Surest. UHG has been touting Surest on its earnings calls for years now as one of the brightest spots in its commercial book, following its acquisition of Bind several years ago. Bind is a great example of what success in this market will look like.

A successful challenger health plan quickly becomes a winning product for BUCA, which then drives reactions from Cigna (Clearity), Aetna (SimplePay), and the Blues (Coupe) to implement a similar product offering. In that sense, it seems like the phrase “the more things change, the more they stay the same” will very much be applicable here. Looking back at the previous eras of generational shifts in plan designs for employers is a worthwhile exercise here, i.e. UHG acquiring Definity in 2004 for $300 million. At the time, Definity was an innovative plan design that would help consumers take control of healthcare spending by introducing higher deductibles tied to medical savings accounts. Today, high deductible health plans are lamented as one of the key consumer issues in healthcare, spurring the next wave of employer plan design innovation.

Speaking of the employer market disruption, I always find it interesting to keep an eye on how the public markets are discussing the potential disruption as helpful grounding for the conversation. With that in mind, it is worth paying attention to Omada’s inaugural investor day, which was held on Thursday.

Omada has seemingly executed about as well as one could hope on the public markets, including telling a narrative around meaningful margin expansion that is ahead of expectations. And as it relates to the disruption front, if the PBM and employer health plan market is at risk of being replaced, you wouldn’t know it from that session. Specifically, this slide below highlights how Omada expects to grow over the next several years via its relationships with the Big 3 PBMs and major health plans:

It’s worth noting here that despite positive analyst commentary coming out of the session, Omada’s stock fell almost 12% on the week. It’s hard for me to put too much weight on stock market swings these days, given it’s unclear how much of it is a reaction to the underlying strategy versus algorithmic trading flows triggered by broader macro conversations, but nonetheless, worth noting.

The Program of All-Inclusive Care for the Elderly, PACE, is one of my favorite topics to discuss in all of healthcare, on so many different levels. At its core, you have a clinical model with a decades-long track record of providing high-quality care for a complex, underserved elderly patient population. At the same time, PACE has never quite scaled, with 204 PACE programs serving only 97,000 participants nationwide today. That’s roughly 475 participants per PACE program. This, despite the fact that PACE, rather controversially, began allowing for-profit programs to participate a decade ago.

Enter InnovAge, the largest PACE provider in the country, which reported its FY Q4 2026 earnings this week. A reminder of just how different this model is: InnovAge expects to have ~8,700 patients in FY 2027, generating ~$1.1 billion in revenue. It’s a public company with a $1.5 billion valuation treating those ~8,700 patients. It employs roughly 2,500 people, which equates to one employee per every 3.5 patients. If you’re excited about the transformational potential of AI in healthcare, I’d suggest that starting at this end of the healthcare needs spectrum is a significantly more meaningful opportunity than in the peptide/wellness world, but that’s just me.

InnovAge’s earnings discussion this week highlighted how InnovAge leadership believes the company is entering a third era, one focused on growth in the PACE program. This quote above highlights the momentum.

Keep in mind that Dr. Oz visited an On Lok clinic in early 2025, RFK has more recently visited InnovAge, and HHS released a study earlier this summer highlighting how PACE models outperform others. That is a lot of attention from DC leadership on a model that currently treats 97,000 people nationwide. Blair suggested that there could be upcoming activity on a couple of fronts: 1. making it easier to enroll patients in PACE; 2. making it easier for PACE programs to expand geographies; and 3. a CMMI model testing expanding PACE into Medicare populations.

The CMMI model would be fascinating to see, particularly because of the underlying logic Blair articulates: expanding PACE into Medicare would potentially save both the federal government and states dollars. In the current macro climate, that seems like a winning proposition. I’d imagine we’ll be talking more about this in 2027.

Head of Clinical Programs & Operations at Sensorum Health, an AI and tech-enabled monitoring and care company for high-risk Medicare populations.
$150k - $170k | Hybrid (NYC)

Greenbrook Medical, a senior-focused primary care provider, is hiring for the following roles:

SpringTide Rising Fellowship at SpringTide Ventures, an early-stage healthcare VC firm.
Unpaid | Remote

Director of Revenue Cycle Management at Pair Team, a community-based primary care platform for Medicaid and Medicare populations.
$185K - $215K | Remote

Director, Medical Economics & Health Outcomes at Atlas Oncology Partners, an oncology-focused VBC company.
$115K - $150K | Remote

Contact us to feature roles in our newsletter.