By CLAUDE.AI trying to write like me MATTHEW HOLT

I finally did what I have been wanting to do for a while and had an AI ingest everything I’ve written on The Health Care Blog. Given there’s 20 years of my posts intermingled in hundreds of guest posts, it wasn’t as easy as I’d hoped. But over three days I finally got to about 500 or so articles I’ve written and asked Claude to ingest my style. I then gave it three decent sized prompts which I’d previously used on linkedin or in a Slack message and it spat out this—Everything below is AI with no edit and no re-prompting. And it is about 60% of my tone and maybe 75% of my opinion. Pretty amazing–Matthew Holt

The Private AI Bubble Meets the Public Digital Health Reality Check

So Garner Health just raised $118 million at a $1.35 billion valuation, and Talkiatry pulled in $210 million. Meanwhile, you could buy both Talkspace (market cap ~$640 million) and BetterHelp—Teladoc’s entire mental health division that once did over a billion in revenue—for under a billion dollars combined given Teladoc’s current market cap of $953 million.

Something doesn’t add up here, and it’s not the math.

We’re watching a fascinating divergence between what private market investors think digital health companies are worth and what public market investors are actually willing to pay for them. And stuck in the middle is the latest gold rush: AI-powered revenue cycle management.

Let’s start with Garner. I like what they’re trying to do—using massive claims datasets to steer people to higher-quality doctors with financial incentives. It’s clever. But 12% in cost savings in the first year? The Health Rosetta crowd claims they can get their members far better results. The direct primary care groups working with employers are reporting bigger numbers without needing a $1.35 billion valuation to do it. And they don’t have a medical group like Included Health backing them up.

So why the massive valuation? Because “AI-powered” and “big data” and “320 million patient records” sound really good in a pitch deck. Never mind that the actual demonstrated savings are less impressive than what consultants have been delivering for a decade.

Then there’s Talkiatry, which just raised $210 million—on top of their previous $130 million Series C. Good for them, they’ve built something real: 800+ employed psychiatrists delivering actual care with decent outcomes data. But here’s what I can’t figure out: why is every mainstream mental health company still raising hundreds of millions in private markets when literally nobody has gotten out?

Talkspace trades at a $640 million market cap. Teladoc owns BetterHelp, which peaked at over $1 billion in revenue and is now in the dumpster, dragging down Teladoc’s market cap to under $1 billion total. Where’s the Lyra IPO? The Brightline exit? The Spring Health or Headspace liquidity event? They don’t exist.

The public markets are basically screaming: “We don’t believe digital health companies are worth what VCs think they’re worth.” And yet the private money keeps flowing.

Now let’s talk about where the really stupid money is going: RCM AI. We’ve got Anterior raising $64 million total to help payers process prior authorizations faster, and an entire ecosystem of companies raising massive rounds to help providers fight those denials. As I’ve written elsewhere, we’ve gone from the smartest people in the world getting consumers to click on ads to half of them trying to squeeze 5% more revenue out of payers for providers, while the other half are building AI to help payers stop them.

This is the dumbest arms race in healthcare technology. We’re automating the negotiation over the crumbs instead of fixing the system that creates the waste in the first place. But VCs love it because there’s a clear value proposition: “We’ll get you an extra $X million in reimbursements” or “We’ll save you $Y million in denied claims.” Never mind that we’re spending billions to build technology that makes a fundamentally broken payment system slightly more efficient at being broken.

Meanwhile, back in the public markets, actual digital health companies with actual revenue and actual customers are trading at valuations that suggest investors think they’re going to zero. Teladoc’s market cap is $953 million. They did $2.53 billion in revenue last year. That’s a 0.37x price-to-sales ratio. BetterHelp alone was doing over a billion before the bottom fell out.

The divergence is getting absurd. Private companies with unproven business models and uncertain paths to profitability are raising at sky-high valuations. Public companies with real revenue, real customers, and real operations are trading like they’re distressed assets.

So what’s happening here? A few things:

First, private market investors are betting on potential while public market investors are pricing in reality. Potential sounds better in a conference room than on an earnings call.

Second, we’re in an AI bubble. Slap “AI-powered” on anything healthcare-related and suddenly you can raise at multiples that make no sense. Doesn’t matter if the AI is actually doing something useful or just glorified automation of existing processes.

Third, the exit window for digital health has been effectively closed for years. The SPAC boom and bust taught us that taking half-baked digital health companies public ends badly. So private companies keep raising private money because that’s the only money available, creating a massive valuation disconnect.

Fourth, and this is the one nobody wants to talk about: a lot of these companies might not be worth what VCs are paying for them. The mental health companies keep raising because they can, not because they’ve proven a path to sustainable profitability. The RCM companies are raising on the promise of extracting value from a broken system rather than fixing it.

Here’s my prediction: We’re going to see a reckoning. Either these private valuations are going to come down to meet public market reality, or we’re going to see a wave of down rounds and shutdowns. The public markets aren’t wrong about digital health—they’re appropriately skeptical of companies that haven’t proven they can make money at scale.

And the AI RCM gold rush? That’s going to end when someone realizes we’re spending billions to automate arguing over money instead of spending billions to reduce the total cost of care. Because at the end of the day, every dollar these companies “save” or “recover” is a dollar that shouldn’t have been in dispute in the first place.

The smartest investors should be looking at the public market valuations and asking: if actual digital health companies with real revenue can’t get respect from public investors, why are we valuing private companies like they’re going to be worth 10x more when they finally try to exit?

Unless, of course, nobody’s planning on exits anymore. In which case, we’re not building companies—we’re building very expensive science projects funded by LPs who apparently have infinite patience and capital to burn.

I’ve been covering this industry long enough to know how this movie ends. And spoiler alert: it’s not with a bunch of unicorn IPOs.

Claude is a clever chatbot made by Anthropic. Matthew Holt hopes it replaces him soon