Roundup: Dollars, Distribution, and the Deals Shaping Digital Health

A word from Second Opinion’s Christina Farr 

It has been one of those heavy travel periods, but wonderful to make stops in Houston, Nashville, Boston, and Chicago. After spending several weeks in the company of health system executives, one thing felt abundantly clear: There’s massive concern about workforce shortages plaguing healthcare. And those shortages, which we face across physicians, nurses, and other providers, are about to get worse as our population ages and a whole generation starts to retire.

There is massive enthusiasm about AI — but also with strong self-awareness that health systems aren’t set up for quick, easy adoption of new technologies. I met with executives in charge of implementation processes who noted that these timelines will not be fast – one year to sell and pilot, one to two years for implementation, and then real investment is needed to ensure the technology is being used appropriately. Investors looking to make a quick buck by chasing AI because it’s a shiny object will face a rude awakening. 

Meanwhile, in the employer landscape, the sense of impending doom seems to be very real. Watch this space for upcoming features on the cost crisis, as employers face the dual pressures of rising hospital costs and specialty drug costs. Employees may be seeing their salaries go up but take-home pay is stable because of the increasing amounts going straight into healthcare they may or may not use. Hard choices will have to be made. I just heard a rumor about a Fortune 100 employer that said “no” to covering GLP-1s – and as word spreads, others will follow suit. 

With crisis comes opportunity. If we look back into history, it’s only in moments where it seems like it can’t get any worse that we see real creativity in healthcare. Reference-based pricing is just one example. I’m genuinely excited to see what comes from this particular moment. Sometimes it feels in healthcare that all the incentives are stacked up to make it impossible to do the right thing. But bad incentives that raise prices can’t exist in perpetuity. There’s only so much the system can take, and if we’re not already at the limit — we’re about to surpass it. 

Why AI adoption will be different from past tech waves?

JB: Health systems were late to cloud because integration was complex, ROI was uncertain, and regulatory risk was high. AI is different–it addresses urgent pain points such as labor shortages, reimbursement pressure, and documentation burdens. The value is tangible in weeks, not years. And, most importantly, AI's interface is plain language. Unlike past technologies that required coding or specialized training, anyone, whether they be technical or nontechnical, can interact with AI naturally. This removes barriers to entry, accelerates adoption across all roles, and makes AI more intuitive than any prior technology wave.

Additionally, healthcare data is uniquely multidimensional, spanning clinical records, imaging, audio, genomic data, and more. This makes it fertile ground for multimodal AI applications that can create a comprehensive view of patient health, from the molecular level to entire populations.

And as more use cases scale, the winners won't just be those with good models; they'll be those who control the key leverage points in the healthcare stack: where data originates, how workflows connect, and how integrations create network effects across systems. Architecture and distribution matter as much as the underlying model.

Biggest regulatory change needed?

JB: The most important change is clear frameworks for the safe, governed use of patient data in AI training and deployment. Current HIPAA-era rules, designed for static data storage rather than dynamic model training, leave providers hesitant. We also need standardized pathways for clinical validation, and clear liability frameworks for AI-assisted decisions.

Equally critical is reimbursement: if health systems and clinicians cannot get paid for AI-enabled tools, whether through CPT codes or bundled payments, adoption will stall regardless of clinical value.

The industry faces urgent economic sustainability challenges, with hospitals struggling financially and life sciences seeing diminishing R&D returns. In this context, adaptive regulation is not optional. It is an imperative.

Finally, regulation must also extend upstream to data preparation and normalization. Many AI failures occur not in the model itself, but in how raw data is cleaned, structured, and converted into signals. Clear accountability for these "conversion layers" will be essential.

Which LLM is best suited to healthcare and why?

JB: It depends on which part of the market you are serving. Enterprise healthcare will favor models that emphasize trust, safety, and compliance over raw scale. I'm particularly excited about Anthropic's Constitutional AI approach. By embedding safety principles during training (rather than filtering after the fact), Anthropic provides governance, liability protection, and audit trails that align with FDA guidance and healthcare's ‘do no harm’ culture. OpenAI has driven mass-market adoption through ChatGPT and is advancing strong reasoning and multimodal capabilities. Google has deep healthcare research credentials with Med-PaLM 2 and partnerships with AMCs like Mayo Clinic.

The landscape is evolving rapidly, but one thing is clear: having a strong model is not enough. The value accrues to those who own key points of integration and distribution. That's where true defensibility lies.

Who will win in AI + healthcare?

JB: The winners will be those who partner effectively – AI companies bringing technology, agility, funding, and compliance, and healthcare incumbents bringing clinical expertise, trust, data, and distribution.

In the long-term, models alone won't win. The combination of a technical moat with strategic control of distribution and integration points is what will separate the winners from the rest. Specifically, winners will position themselves at critical junctures where data originates (e.g., patient-provider conversations, prescription decisions, or clinical trial protocols) creating cascading value across multiple stakeholders and enabling network effects that are difficult to replicate.

With Annalisa Merelli 

The news: Millie, a hybrid clinic based in California currently focused on preconception and maternity care, will begin offering treatment for perimenopause and menopause.

What Millie said: “There is a population that is sort of postpartum and not quite in menopause that is kind of falling through the cracks as they navigate that transition,” Anu Sharma, founder and CEO of Millie, told Fierce Healthcare. 

Perimenopause is getting the attention it deserves: Perimenopause is an increasing focus for women’s health companies, as it often hits women at the peak of their earning potential. Many women are still growing their families as they hit perimenopause, ensuring it dovetails with the fertility industry, and there’s an increasing number of patients prescribed a combination of hormone therapies and GLP-1s. 

The news: The Covid waivers that allowed Medicare and Medicaid patients to access telehealth visits (for visits conducted outside designated clinical sites) expired on October 1, and some hospitals and health centers are informing patients that they can no longer access remote visits. 

The prognosis: The waivers could be renewed when the shutdown is resolved, though it is not a given that providers that chose not to discontinue remote visits will be compensated retroactively. 

Why it matters: Though this will eventually be resolved, it does give a measure of just how much, and how quickly, the US healthcare system has come to rely on telehealth. Telehealth didn’t see mass adoption for decades because of the lack of consumer and provider awareness. Because of the pandemic, that’s no longer the case. There are powerful lobbies that will continue to push for reimbursement and access, because patients are demonstrating they value convenience. 

The news: Telehealth has unlocked the age of being licensed in all states. Doctors who accumulate licenses are extremely valuable, because they can be owners of telehealth companies and serve as medical directors supervising practitioners across states. 

The numbers: In 2016, only nine doctors were licensed to practice in all states. In 2024, 172 were licensed in all 50 states, and 356 had acquired at least 45 licenses. 

Why it matters: This highlights the challenges of adapting the current in-person regulation system to telehealth. As virtual care becomes more and more a prescription-first field — as Katie Palmer details for STAT — physicians may have more success because of their licensing status rather than their actual skills. Second Opinion’s very own Christina Farr was the first to cover this trend at CNBC back in 2019, when it was still a rarity given the expense and administrative hassle required to get all 50 licenses. 

The news: According to an analysis by PwC, about $1 trillion will shift from healthcare spending in brick-and-mortar facilities with high overhead and complex administrative burdens toward AI-enabled, digital-first healthcare offerings. 

The drivers: The analysis says several factors will drive the shift — a change in chronic condition healthcare that moves away from management and toward prevention; greater innovation in remote care offerings; the emergence of a very informed patient/consumer with higher disposable income and greater agency in choosing their care. 

One crazy number: The total US healthcare spending for 2035 is projected to be $8.6 trillion.

Why it matters: Digital health skeptics will point out that there haven’t been many successful exits yet in the industry, despite all the investment flooding into the sector. But those who remain bullish on its prospects — ourselves included — will argue that it’s still early days for the sector. And healthcare is a notoriously slow and challenging segment to penetrate, given there are powerful incumbents and misaligned incentives. Watch this space. 

The news: Amazon Pharmacy will install prescription-dispensing kiosks in One Medical locations, starting in Los Angeles, so that patients can get their medications immediatley after their visit.

What Amazon said: “What we’re doing is starting from the customer and working backwards,” Hannah McClellan, vice president of operations at Amazon Pharmacy, told Fierce Healthcare. 

Why it matters: Amazon isn’t the first to explore kiosks for convenient dispensing, but it’s one of the best positioned companies to pull it off. Given that there’s already One Medical locations in many cities across the country, and nearby jumbo employers, all it takes is the additional step of adding on-site dispensing capabilities. Where this could really add value is with adherence, particularly for patients juggling many different medications or those who don’t live nearby a pharmacy. 

The news: On Friday, after announcing a deal with AstraZeneca to lower drug prices (following a similar deal with Pfizer last week) in exchange for a reprieve from tariffs, the administration promoted TrumpRX.org, a portal — coming soon — that would help people to buy drugs directly at a low price, bypassing insurance. 

What Trump said: “AstraZeneca will also list many of their most popular drugs online at Trumprx.gov. Trump RX.I don't know why they put the name Trump. I did not tell them to do it, but I'm honored to let him do it.”

The catch: Donald Trump, Jr. sits on the board of BlinkRx, an online direct-to-consumer drug delivery company that stands to benefit from TrumpRX, which would redirect consumers to such DTC sites, the Wall Street Journal reported.

If you have been following this newsletter, you’ve noticed it’s been a busy time for health tech deals. STAT’s Mario Aguilar puts it into some context, explaining how larger companies are M&Aing their way to innovation and development by acquiring smaller players.  

DUOS secures $130 million: The AI-powered platform aims to engage patients, particularly older adults, and guide them through their benefit offers. It received a strategic growth equity investment led by FTV Capital.

HealthTap joins LillyDirect: The virtual care company is the latest to join the roster of providers on Lilly’s direct-to-consumer care site. It will be available for the treatment of Type 1 and Type 2 diabetes. The financial details of the partnership were not disclosed. 

$29 million to expand DreemHealth: The funding, led by Eurazeo, aims to expand the digital sleep-health clinic, making it available in all 50 states. 

Heidi Health raises $65 million: The AI medical scribe company closed a Series B led by Steve Cohen’s Point72 Private Investments and announces a new AI tool that calls patients on behalf of doctors. 

The Medical Travel Company raises $4.5 million (€3.8 million): The India-UK startup on a quest to streamline medical tourism received seed funding led by Nexus Venture Partners.

A $6.75 billion AI deal for Qualtrics: The AI-powered customer survey company agreed to buy Press Ganey Forsta, a healthcare research company, the Financial Times reported. The deal would give Qualtrics access to the research company’s database and network. 

Nava Benefits gets $30 million: The benefits brokerage tech company closed a Series C funding round led by Thrive Capital.

Attuned Intelligence raises $13 million: The supervised voice-AI, call center system designed for healthcare closed a seed funding round led by Radical Ventures and Threshold Ventures.

Sensi.Ai raises $45 million: The senior care tech company closed a Series C funding round led by Qumra Capital, bringing its total funding to $98 million.