Tele-Doom: Why AI Is Rewriting the Future of Telehealth

The telehealth boom was supposed to revolutionize healthcare forever—but what went wrong? In this episode, Dan McCoy unpacks the dramatic fall of industry giants like Teladoc and Amwell, revealing how their high-profile bets on nationwide distribution networks failed to stand the test of time. More importantly, you'll hear why the real disruptor isn’t a return to in-person care, but the explosive rise of AI-powered tools that are decentralizing healthcare delivery.

Dan breaks down the structural shifts pushing telemedicine incumbents to the brink, explores the rapid adoption of ambient clinical AI, and explains how local practices now leverage advanced technology to deliver more personalized, context-rich care. If you care about the future of healthcare—from benefit managers to health system execs to curious entrepreneurs—this episode is your essential guide to what’s next, who’s winning, and why yesterday’s telehealth playbook no longer applies.

Check out my SubStack for a deeper analysis: 

 https://danmccoymd.substack.com/

Transcript

SPEAKER_00 0:07

Here's something that should make every healthcare investor uncomfortable. In February 2021, Teledoc Health peaked at around$294 a share. The future of medicine, we were told, the pandemic had proven that telehealth was inevitable. And Teledoc was so confident it spent$18.5 billion acquiring Lavongo, a chronic disease management platform. Today, Teledoc trades in the mid-single digits. That's roughly a 95 to 98% decline, depending on the day. They wrote off$13.7 billion of that Lavongo acquisition in a single year. Amwell, their closest competitor, tells a similar story, down more than 90% from its early 2021 peak, trading in a market cap smaller than some Series B startups. Now the easy explanation is that this was just a pandemic bubble that popped, and that's partly true, but it's also incomplete and dangerously misleading, because what's actually killing traditional telemedicine isn't the return to in-person visits. It's something far more structural. I'm calling it Teledom, and it has almost nothing to do with the pandemic hangover. To understand what's really happening, you need to think about what telehealth companies actually built, because it wasn't technology, it was distribution. Teledoc, Amwell, MDLive, their core value proposition was never the video call itself. Any developer could build a video platform. What they built was a nationwide network of licensed physicians integrated with insurance panels, credentialed across state lines, and plugged into employer benefit packages. That was the moat. That was the$18.5 billion bet. The model, build a massive national distribution network, connect patients to anonymous rotating physicians, was modeled on the same logic as Uber. Scale the supply side, own the demand side, and win through network effects. And for a few years, particularly 2020 and 2021, it appeared to work. Telehealth usage hit roughly a third of all medical visits in Q2 of 2020. But by 2023, it had cratered to mid-single digits. Trillion Health estimates about 6%. Optim owned United Health Group, the largest healthcare company in the world, and they launched its virtual care business in 2021. They shut it down in April of 2024. Walmart launched telehealth the same year, also shut down the same month, citing, and I'm quoting here, the challenging reimbursement environment at escalating operating cost. So the largest health insurer and the largest retailer in America both tried national telehealth distribution and walked away. That should tell you something. The conventional wisdom is straightforward. People prefer seeing their own doctor in person. The pandemic forced telehealth adoption, the pandemic ended, and utilization returned to baseline. Teledoc's problems in this telling is cyclical demand, not structural obsolescence. And this story has some evidence behind it. The CDC's National Health Statistics Report show that telehealth stabilized at around 4.5% of visits after 2022. That's still meaningfully higher than the less than 1% pre-pandemic baseline. A Kaiser Permanente study found that patient satisfaction with telehealth was genuinely high, around 85%, when patients saw their own doctor physician virtually. The problem was with the Teledoc model, specifically, random physicians, no continuity and no context. If you're a Teledoc bull and there are a few left, you argue that the company is pivoting to integrated care, that BetterHelp provides a mental health moat, that chronic care management through Lobongo still has legs. Here's the problem. BetterHelp revenue declined around 8% in 2024 and continued falling in 2025. The integrated care segment grew low single digits annually, roughly the rate of healthcare inflation. In other words, Teledoc's growth business is growing at the speed of doing nothing, and their consumer business is actively shrinking. But the stock price isn't the story. The stock price is a symptom. The story is what's replacing them. In 2025, U.S. digital health startups raised on the order of$14 billion. That's the highest since 2022, according to Rock Health. Here's the critical detail. A majority of that funding, over half, went to AI-enabled companies, and that's up from about a third the year before. AI-enabled healthcare startups raise significantly more per round, an 80% premium over non-AI startups. And the category seeing the most explosive growth isn't telehealth, it's ambient clinical AI. Analysts estimates this market generated roughly$600 million in revenue alone in 2025, a 2.4x increase year over year. By June 2025, 63% of Epic hospitals had deployed ambient AI documentation, and that's up from essentially zero just two years earlier. That's one of the fastest technology adoption curves in healthcare IT history. A bridge, a startup most people haven't heard of, raised over$500 billion across two recent rounds, reaching a$5.3 billion valuation. They're now deployed across 150 health systems, including Kaiser Permanente's 40 hospitals and 600 medical offices. That Kaiser deployment was their fastest technology implementation in over 20 years. Hippocratic AI hit unicorn status in January 2025 with a$1.6 billion valuation. Their product, AI agents that handle patient calls, care navigation, follow-ups, and scheduling. The exact functions that Teledoc built a nationwide physician network to perform. And here's where this matters most. 85% of all generative AI spend in healthcare is going to startups, not incumbents. Even Microsoft's Nuance, deployed in 77% of U.S. hospitals, is losing market share to smaller, more nimble AI companies. Startups like Abridge and Ambiance have clearly captured nearly 70% of the new ambient AI market, despite Nuance's massive installed base. This brings me to the real thesis. And it's not just that AI is replacing telehealth, it's about where the intelligence lives. Traditional telemedicine was built on a centralized model. You, the patient, connect through a national platform to an available physician who knows nothing about you. No access to your records, no relationship with your local hospital, no understanding of your pharmacy, your specialist, your history. The physician reads from your intake form, asks generic questions, and makes a best guess. This is what passes for innovation healthcare, a doctor visit, but worse, and on your phone. Context-aware AI flips this entirely. The new model doesn't connect you to a stranger. It connects you to your own physician's practice, even a small one, to AI systems that are deeply integrated with your electronic healthcare records, your local hospital network, your pharmacy data, your wearable devices, and your care history. A study published in Nature in February 2026 tested AI chatbots in clinical settings in Rwanda and Pakistan. In Rwanda, the chatbot answers outscored those local clinicians across every metric assessed. In Pakistan, physicians using large language models achieved a mean diagnostic reasoning score of 71%, compared with just 43% using conventional resources. As Caroline Green, Director of Research at Oxford Institute for Ethics and AI, put it, these tools might be able to support clinicians in lower and middle-income countries to improve the level of care. But here's what the Pakistan study actually proved that most people missed. The AI wasn't replacing the physician, it was making the local physician dramatically better by providing contextual decision support. The advantage wasn't the model, it was the context the model had access to. This is the decentralization that is already undermining Teladox model. A solo practitioner's office in rural Texas can now deploy an AI system that handles patient triage, drafts clinical notes from the visit conversation, manages follow-up scheduling, and provides real-time clinical decision support, all integrated with that patient's actual medical records. A majority of smaller practices surveyed now are expect to be using AI automation within two years, and they don't need Teledoc's national distribution network to do it. They need a software subscription and an EHR integration. The competitive advantage has shifted from who has the biggest physician network to who has the deepest contextual understanding of the patient. And on that metric, your local doctor with an AI co-pilot beats a random Teledoc physician every single time. So what does this mean if you're an employer buying benefits or health system executive or frankly anyone making decisions about healthcare delivery? First, and this matters for every employer reading the renewal quote, the telehealth line item in your benefits package is increasingly hard to justify. You're paying for a panel of anonymous physicians when an AI augmented version of your employee's own doctor will be more contextually informed. Ask your benefits consultant what their telehealth utilization rate actually is, and I'm betting it's below 5%. Second, if you're running a healthcare system, the question isn't whether to adopt AI. 63% of Epic hospitals already have. The question is whether your AI strategy is a bolt-on feature or a fundamental redesign of how patients interact with your system. Hippocratic's AI Agent App Store, that's where clinicians design custom AI agents for their own specific workflows, it kind of gives you a preview of where this is all heading. Every practice becomes its own AI-enabled healthcare platform. Third, and this is the uncomfortable one, Teledoc and Amwell still have significant enterprise contracts. These companies are finding a structural decline with incremental improvements. Teledoc's 2026 revenue guidance,$2.5 billion to$2.6 billion, is essentially flat to down. Their adjusted EBITDA guidance suggests they're managing for cash preservation and not growth. The mainstream narrative about telehealth's decline is that the pandemic created artificial demand, now we're back to normal. That's comfortable, but it's also wrong. What actually happened is more interesting. The pandemic didn't just inflate telehealth demand temporarily, it accelerated a deeper shift in how healthcare technology gets distributed and deployed. The companies that will define the next decade of healthcare are building a bigger physician network. They're building contextual intelligence layers that make every physician in every practice in every town dramatically more capable. The distribution advantage that Teledoc spent billions to acquire is rapidly being eroded by a five-figure software subscription. The question isn't whether AI replaces telemedicine, it's whether the telemedicine incumbents can become AI companies fast enough to survive. If you found this useful, hit subscribe and the notification bell so you don't miss future episodes. I go deeper on topics like this over on my Substack, links in the description, where I'll share the data sources and additional analysis that didn't make it into this video. Until next time, I'll see you then.