Sponsored by The RN Network Job Board
Reduce Nursing Time to Fill — unlimited nurse postings, one low price.
Receive unlimited LPN, RN, Nurse Leadership, and Advanced Practice nursing roles on The RN Network job board. Get found in search results, mentioned on Nursing Matters podcast, in nursing email blasts, and across The RN Network’s social media channels.
Buy 12 months, and receive a free six month subscription to Workforce Intelligence, weekly hiring benchmarks and insights. Offer good through 10/31/2026.
Learn more and register today →The RN Network job board is our own, so no commission changes hands this week — the sponsor slot is house-run.
The short version
- The shortage figure in nearly every citation — 267,330 full-time RNs by 2028 — comes from HRSA’s December 2025 projection and sits one year past that file’s own peak.
- The national gap bottoms at 267,900 full-time equivalents in 2027, then narrows in every single year after that, to 108,960 by 2038.
- HRSA’s model has no technology term in it. The gap closes on licensure and labor force participation alone.
- Published AI capacity results are nowhere near national-gap scale. A real named program returns single-digit full-time equivalents a year.
- Of the 36 jurisdictions ever in deficit, 23 hit their worst year in 2027 or earlier. Seven do not reach it until 2035 or later.
Somewhere in your board deck is a slide with the nursing shortage on it. The figure on that slide is almost certainly 267,330, and it almost certainly carries the year 2028.
Open the file that number came from and look at 2029.
The shortage figure in every citation you have read is one year past its own peak
Start with where the number comes from. HRSA published its Nurse Workforce Projections in December 2025, built on a 2023 base year and running through 2038. The Health Workforce Simulation Model behind it reports projected supply, projected demand and the gap between them for every year and every state. The American Association of Colleges of Nursing quotes the 2028 line on its nursing shortage fact sheet, updated this month, as a projected shortage of 267,330 full-time RNs. That is the figure in circulation, and it is accurate.
Here is the rest of the column it sits in. The national RN gap widens from 183,110 full-time equivalents in 2023 to a low point of 267,900 in 2027, holds almost flat at 267,330 in 2028, and then narrows every single year after that — 250,970 by 2030, 186,980 by 2034, 108,960 by 2038. Adequacy of supply against demand bottoms out at 91.8% and recovers to 97.0%.
The worst year in that file is next year.
The turn is not uniform, and that part matters more than the national figure does. Thirty-six jurisdictions are in deficit at some point across the horizon. For twenty-three of them the single worst year is 2027 or earlier. Seven do not reach their worst year until 2035 or later. Your market has a peak and it is probably not the one on the slide.
My read: the 2028 number is the one in circulation because it is the largest one anybody has a reason to quote. Nearly everyone citing a shortage figure is citing it in support of something — a budget, a bill, a product, a headline — and a curve that improves supports none of those.
The model closes that gap without a single AI assumption in it
Look at what the forecast is actually made of. HRSA converts projected demand into workers using staffing ratios, and bases hours worked and retirement patterns on 2019 through 2023 data. Its own factsheet states the assumption plainly, that historical patterns of attrition, graduation and labor force participation remain the same. Its technical documentation lists changes in medicine and technology among the uncertainties the projections face, rather than among the things the base case models.
So the curve that closes is a headcount curve and nothing else. More people get licensed, fewer leave, demand growth slows as the population bulge works through. No unit in that model delivers care any differently in 2038 than it did in 2023.
That cuts the opposite way from how a capacity program is usually sold to you. The pitch is that the shortage is permanent and worsening, so you need the technology. The forecast does not say the shortage is permanent and worsening. It says the shortage is at its worst right now, and that it closes on licensure alone given fifteen years.
Fifteen years is not a hiring horizon. Five quarters is.
Put the published AI results in the same denomination as the gap, and the scale problem is obvious
MD Anderson Cancer Center ran virtual nurses through 12,441 patient visits between September 2024 and April 2025 and reported 2,247 bedside hours returned to its nurses. Annualize that and divide it by a full-time year, and the program is worth about 1.6 full-time equivalents. Ochsner Health, roughly three years in, covers 615 beds across 14 hospitals and 19 units, and reports about 7,500 bedside hours saved since expanding in 2024 — about 3.6 full-time equivalents. Both come from Becker’s Hospital Review’s December 29, 2025 compilation.
The national gap at its peak is 267,900.
Nothing in that comparison argues against either program. It argues against one specific sentence, which is the sentence a vendor will put in front of your CFO this quarter — that AI is how the nursing shortage gets solved. At published scale it is not, and it is not close. What these programs do is move one hospital’s position, on specific units, in the years when the market is tightest.
The strongest nursing-specific figure in this space is still the Bon Secours Mercy Health analysis published in Nurse Leader in August 2026, which found 20.6 recoverable minutes per twelve-hour shift at 60% ambient documentation capture or better, across 29,467 Epic encounters on five medical-surgical units at four hospitals. That is peer-reviewed, it is nursing rather than physician data, and it is a per-shift figure.
A per-shift figure is not a headcount figure. Hours returned and positions avoided are two different numbers that do two different jobs in a budget, and conflating them is how a hiring leader loses an argument they should win. The hours are real. Whether they become money depends on whether your own balance sheet has contract labor to stop buying or premium pay to stop paying.
The Workforce Intelligence Center carries this projection by state and target year, so you can pull your own market’s peak rather than planning against the national one.
No new graduate can reach you before the peak, and the teaching bottleneck is why that will not change
Here is the constraint that makes the timing decisive. A nurse who enrolled this fall does not appear on your floor in 2027. Every unit of relief that reaches you inside the peak is either an hour returned on a unit you already staff, or a nurse taken from somebody else’s.
The supply side of HRSA’s curve is an assumption rather than a delivery. IPEDS counted 228,331 associate and baccalaureate completions in registered nursing across the fifty states and the District of Columbia for academic year 2024, from roughly 1,110 and 1,160 institutions respectively. AACN reports 92,672 qualified applications turned away from baccalaureate and graduate programs in 2025, nearly 17,000 of them graduate applications. Its October 2025 survey of 1,075 nursing schools, 863 of which responded, found a 7.2% nurse faculty vacancy rate and 1,588 unfilled full-time faculty positions, 80.9% of them requiring or preferring a doctorate — while PhD enrollment fell for an eleventh consecutive year.
Entry-level baccalaureate enrollment did rise 7.6% in 2025, which is the good news in the file and the reason the curve bends at all.
The curve that closes HRSA’s gap runs through a classroom that cannot get bigger without faculty the system is not producing.
Five moves before your 2027 sourcing plan is signed
Date your sourcing plan to your own market’s peak, not the national one. Pull the peak year for every state you recruit in. Twenty-three of the thirty-six deficit jurisdictions have already hit their worst year or hit it in 2027; seven do not reach it until 2035 or later. One sourcing timeline applied across both groups will be wrong in most of your markets.
Stop writing “worsening nursing shortage” into any budget request that runs past 2028. A CFO who opens the same HRSA file sees the gap shrinking in every year after 2027. Ask for the money against the peak you can show, which is immediate, rather than against a decade the source does not support.
Ask your operations counterparts what capacity work is in flight before you commit next year’s requisition plan. Hours returned on a unit change how many lines you are asked to fill, and that decision reaches your plan without passing through your office first. Find out now which units have a program live and which are in pilot.
Price retention against the peak rather than against the decade. The nurse you hold through 2027 and 2028 is the one you cannot buy at any reasonable number, since every competitor in your market is at their own tightest point in the same two years. After that the forecast describes a looser market.
Staffing firms: sell the twenty-four, not the fifty-one. Twenty-four jurisdictions are still in deficit in 2038 and the rest cross into surplus before then. Multi-year contract demand lives in the first group, and California alone is projected 84,750 full-time equivalents short at 78.1% adequacy in 2038.
You are being asked to defend a number whose own author has it shrinking
Every hiring leader I talk to has one slide with the shortage on it, and almost none of them have read past the year on that slide. It is not carelessness. The figure arrives pre-quoted, from a credible source, in support of a request they genuinely need to make, and there is no reason to open the file behind it.
The reason to open it is that the file contains a date, and the date is the part you can act on. The national gap is at its worst between now and roughly fourteen months from now. That window closes whether or not anyone budgets against it, and no amount of recruiting investment produces a nurse inside it who was not already licensed.
Pull your own states’ peak years this week, and put the two numbers in front of your finance partner together — the year your market is tightest, and the year it loosens. Build the ask around the first one.