602 Hospitals Are Already Underwater. Here's What CFOs Need to Do Before the Next Wave Hits.

I spent this week's LinkedIn newsletter on AI oversight, or the lack of it. This week's blog post is about money, because if you run a hospital, a health system, or an RCM vendor that serves either one, the money is about to get a lot scarier than the algorithm.

On July 6, National Nurses United published a report called "A Preventable Crisis." The headline number: 602 U.S. hospitals could see their combined deficit grow by 50 to 75 percent once three federal revenue cuts finish landing. I read the whole thing, appendices and all. Here's what's actually in it, what's advocacy framing versus hard data, and what I think you should be doing about it before your board meeting turns into a fire drill.

What Did the Report Actually Find?

NNU screened 3,911 hospitals using five years of Medicare Cost Report data through the end of 2025, and scored each one on a Financial Vulnerability Index built off net income and net worth trends. Anything that scored -7 to -10 out of a possible -10 got labeled financially vulnerable. That cutoff caught 602 hospitals.

Those 602 already carry a combined $10.16 billion deficit, and that number predates a single dollar of the three cuts NNU modeled: the 2 percent Medicare sequestration, the Medicaid funding reduction tied to HR 1, and the expiration of the enhanced ACA marketplace subsidies. Layer all three on top of the existing deficit and NNU projects an additional $5.21 billion to $7.72 billion in losses, bringing the total projected shortfall to somewhere between $15.37 billion and $17.88 billion.

The median hospital in that cohort would lose $3.4 million to $4.7 million. A handful of the largest would lose up to $316 million. And 96 of the 602 hospitals had actually clawed their way to positive net income in their most recent reporting year. Under NNU's scenarios, that number falls to somewhere between 23 and 37. Read that again. Fifty-nine to seventy-three hospitals that were finally turning a corner get shoved back into the red, not because they did anything wrong, but because the floor moved.

Is This Just a Union Talking Point?

I want to be straight with you about where this data comes from, because I'd be doing you a disservice if I didn't. NNU represents more than 225,000 registered nurses, and the union has been public and vocal in its opposition to the Medicaid and ACA provisions in HR 1. This report exists to make a case, and the recommendations at the end (cancel the cuts, build a new 101-percent-of-cost reimbursement program for distressed hospitals, have CMS publish this index annually) reflect that advocacy position. Becker's Hospital Review flagged the same thing when they covered it.

Here's the part that matters more than who published it: the underlying financial data comes straight from CMS Medicare Cost Reports, the same self-reported filings every hospital in the country already submits. NNU built a scoring methodology on top of real numbers. You can disagree with their policy prescription and still trust the arithmetic. I'd encourage you to read Appendix B, where they walk through their managed care correction methodology using California and Minnesota as proxies. It's more rigorous than most advocacy reports bother to be, and it's also a reasonable thing to push back on if your own hospital's payer mix doesn't look like California's.

Whatever side of the HR 1 debate you land on, and I know good people on both sides of it, the operational reality doesn't care about your politics. A hospital running a deficit today does not get to opt out of a 2 percent Medicare sequestration or a 10 to 18 percent Medicaid reduction because it disagrees with the policy. It has to plan around it.

The ACA Piece Makes This Worse, Right Now, Not Someday

This isn't a future problem. CMS's own enrollment data, published in late June, shows ACA marketplace sign-ups fell in 49 of 50 states plus D.C. between February 2025 and February 2026. Total enrollment dropped from nearly 21.8 million to about 19.2 million, a 12 percent decline. Ohio, Oklahoma, and Arizona each lost roughly 30 percent of their enrollees. New Mexico was the only state to gain, because it's the only state that fully replaced the expired federal subsidy out of its own pocket.

And 2027 is set up to be worse. KFF's analysis of preliminary rate filings from 77 insurers across 16 states shows a median proposed premium increase of 14 percent for next year, on top of what's already happened. Insurers are naming the same causes over and over: rising underlying medical costs, broader inflation, and the expired enhanced premium tax credits. Nine insurers have already announced they're exiting the exchanges entirely after 2026.

Every one of those newly uninsured or priced-out patients doesn't disappear. They show up later, sicker, in your ED, and they show up as bad debt on your books instead of a clean claim.

What This Actually Looks Like Inside a Hospital Right Now

This isn't theoretical for the CIOs already living it. Stephanie Hines, CIO at Valleywise Health in Phoenix, a safety-net system where roughly half of revenue is tied to Medicaid or the supplemental funding that supports its teaching program, told Becker's her team can't make a single technology decision anymore without a financial case attached. She's betting on reducing administrative burden on clinicians, not because it's a nice quality-of-life initiative, but because every hour a nurse spends on documentation is an hour she can't spend absorbing the patient volume Hines expects to show up as coverage erodes.

Ray Lowe, CIO at AltaMed, the country's largest federally qualified health center, is running the same calculation from a different angle. AltaMed serves more than 500,000 patients, most of them non-English-speaking and at or below 200 percent of the federal poverty level. California's Medicaid expansion used to cover patients regardless of immigration status; new federal eligibility rules are about to knock a chunk of that population out, and AltaMed is chartered to keep treating them anyway. Lowe's answer is deploying AI tools, ambient documentation, note summarization, patient instruction drafting, tied to measurable outcomes, because as he put it, the era of throwing technology at a wall and hoping it sticks is over. Every dollar has to justify itself against the quadruple aim now.

That's the tell. The hospitals with a real plan are the ones treating every technology dollar like it has to earn its keep against a specific, named financial threat. The ones without a plan are the ones hoping the next AI vendor demo solves a problem they haven't actually defined yet.

So What Do You Actually Do With This?

Here's my forecast, and I've been saying versions of this since I started writing about revenue cycle: by 2030, the hospitals still standing will be the ones that treated the next four years as a margin-recovery sprint, not a policy waiting game. Nobody is coming to save you on a legislative timeline you can control. NNU wants the cuts canceled. Maybe Congress does that. Maybe it doesn't. Either way, the hospitals that survive won't be the ones that guessed right about Washington. They'll be the ones that tightened denial prevention, cleaned up coding accuracy, and renegotiated payer contracts while everyone else was arguing on cable news.

A few places to start, based on everything above:

Pull your own Financial Vulnerability Index score. NNU published the full methodology in Appendix A of the report along with the complete list of 602 hospitals in Appendix E. If yours is on it, or close to it, that's your board conversation for next quarter, not next year.

Get current on your state's ACA enrollment and rate trend. If you're in one of the states that lost 20 percent or more of marketplace enrollment, your uncompensated care line is about to move whether or not your budget assumed it would.

Stop treating AI spend and margin recovery as two separate conversations. Valleywise and AltaMed aren't buying AI because it's exciting. They're buying it because it's the only lever left that can absorb more patient volume without adding headcount they can't afford. If your AI roadmap doesn't map to a specific dollar figure you're trying to protect, it's not a roadmap. It's a wish.

I'll be watching what Congress does with the this and HR 1 Medicaid provisions closely, and I'll flag it here the moment there's real movement. In the meantime, hope is not a revenue cycle strategy. Neither is waiting for the machine to sort it out on its own.

Sources: National Nurses United, "A Preventable Crisis" (July 2026); Becker's Hospital Review coverage of the NNU report; KFF analysis of ACA marketplace enrollment data (June 2026) and preliminary 2027 rate filings; Becker's Hospital Review interviews with Valleywise Health and AltaMed leadership.