The Wall the FDA Built, the Price the Market Burned: What's Left of VUNO

In Korean medical AI, VUNO (KOSDAQ 338220) has long been filed under “the ones with substance.” Its flagship product, DeepCARS (VUNO Med-DeepCARS), is software that predicts cardiac arrest risk in hospitalized patients from vital signs — actually deployed in Korean hospitals, actually billed. DeepCARS alone generated roughly 5.3 billion won of revenue in Q1 2026. Competitors openly benchmark VUNO’s non-reimbursed market entry as the “10x jackpot playbook”. This company’s domestic business model became the industry template.

That company’s stock closed July 24 at 7,580 won — a market cap of about 106 billion won. It was 19,230 won on April 24, so more than 60% evaporated in three months; against the 52-week high of 28,350 it is down 73%. What happened can be summarized in one date. April 30. The FDA.

What an NSE letter actually means

On April 30 (local time), VUNO received a “Not Substantially Equivalent” (NSE) determination in the FDA 510(k) review of DeepCARS. The 510(k) pathway clears a device by demonstrating equivalence to an already-cleared predicate; the FDA judged that equivalence unproven.

Nor was this the first delay. Back in January, reports said the FDA was asking for multi-ethnic data, pushing the review out. Does a model trained on Korean hospital data perform equally across America’s demographic mix — a characteristically American regulatory question, and a legitimate one. The company says it will “reorganize the clinical data and promptly refile”.

Worth being precise here: NSE is a rejection, not an expulsion. The refiling path is open and this is not rare in medical AI. But two things are undeniable — US revenue has been pushed back by quarters at minimum, and neither the company nor investors can now date it.

One interesting counter-signal

In the same April the FDA built its wall, another American regulator sent the opposite signal. The Centers for Medicare & Medicaid Services included DeepCARS in its FY2027 NTAP (New Technology Add-on Payment) Proposed Rule — meaning the product’s clinical value and case for supplemental reimbursement are under official review. The licensing agency says not yet; the payer agency says this looks valuable. That is the basis for reading the US story as delayed, not dead.

The numbers

Metric	Value	Note
Share price (Jul 24)	7,580 won	Apr 24: 19,230 → −61%
52-week range	6,770 – 28,350 won	Now at the bottom 3.8%
Market cap	~106bn won
2025 results	Revenue +34.7% · operating loss −60.4%	Was approaching breakeven
Q1 2026 DeepCARS revenue	~5.3bn won	Total revenue −16% QoQ
P/B	~2.94x

Two things stand out. First, the trajectory through 2025 was a textbook turnaround story — 35% revenue growth with losses cut by 60%; brokerages were forecasting a swing to profit during 2026, before the FDA outcome. Second, much of that forecast assumed US momentum, and with Q1 revenue down 16% quarter-on-quarter, whether domestic growth alone can hold the trajectory is now an open question.

A 106 billion won market cap is jarring if you remember how this company was once priced. But “it fell a lot” is not a valuation argument. P/B of 2.9x still prices growth, not assets, and a loss-making growth story consumes cash as time passes.

So — was the drawdown an overreaction?

I don’t give buy/sell advice, and nothing in this piece is a recommendation. Here is what the public record supports.

Reasons to take it seriously. A domestically proven revenue product (5.3bn won a quarter) and a non-reimbursed expansion model the industry benchmarks; the CMS NTAP review as an American value signal; and a P&L that had walked to the edge of breakeven. The stock has round-tripped to a level that effectively zeroes out the US story.

Reasons for skepticism. Both the outcome and the timing of the FDA refiling are unresolved — an application that drew one NSE has no guarantee of passing the second time. The Q1 revenue decline may be a warning about domestic momentum, not just noise. A loss-maker’s time is not free, and clinical work for the refiling costs money. And medical AI as a sector sits outside this year’s theme rotation, so the market may keep testing the valuation floor until flows return.

Checkpoints that actually change the picture. (1) The FDA 510(k) refiling being submitted, with disclosure of the clinical scope — the faster the refiling, the more likely January’s “multi-ethnic data” homework was already in hand; (2) quarterly revenue recovering sequentially — DeepCARS bed-count expansion showing up in numbers; (3) reaching quarterly operating profit. In my personal framework, VUNO is a name whose observation intensity rises the moment (1) lands; until then it stays on the watch list.

Applying the lesson of the Nable case study here: I wrote that in small caps the AI label is a liquidity event, and VUNO shows the reverse also holds — when liquidity exits on one piece of bad news, even companies with substance fall like theme stocks. The difference is what remains afterward. At Nable’s bottom sat the same small business as before. At VUNO’s bottom sit a product doing 5.3 billion won a quarter — and homework with a deadline nobody knows.

Nothing in this article is financial or investment advice. All figures come from public data as of July 24, 2026, may contain errors, and must be verified against primary disclosures before any decision. The author holds no position in any security discussed.

Sources

This article may contain affiliate links. Nothing here is financial or investment advice.