DAC — Danaos Corporation

Danaos Corporation (NYSE: DAC) — deep-value analysis (REFRESH redo)

Analysis date: September 12, 2026 · Reference price: … (close 09/11/2026, yfinance) · Market cap: … on 18,203,567 shares (06/30/2026, Q2/2026 6-K) · Single currency: reporting USD, trading USD — no conversion risk in any calculation in this report.

Regime: REFRESH. The mechanical deep_delta.py triage triggered the redo on a single ground — 1 unverifiable falsifier out of the six from the 08/21/2026 thesis. Five of six falsifiers hold (datorie_capitaluri 0.2984 below the 0.5 threshold; datorie_neta_ebitda 0.3467 below 2; marja_operationala 0.5829/0.6002 above 0.35; pb 0.7137 above 0.4; roe 0.1333 above 0.08). The sixth — fcf_yield < 4% — could not be evaluated at all. This report resolves exactly that point, re-derives the valuation at the current price and explicitly inherits the rest.

What I inherit, unchanged, from the 08/21/2026 report (rapoarte/deep/2026-08-21-deep-DAC.md): the business and moat description, the capital allocation history, the five-quarter Quality of Earnings analysis, the CEO profile and the accounting red flags inventory. The reason is verifiable: there is no new SEC filing after 08/04/2026 (Q2/2026 6-K, Exhibit 99.1), and the next report is scheduled for 11/16/2026 (tracker, urmatoarea_raportare field). The last document in the SEC-Filings\DAC folder is the same 08/04 6-K the reference analysis used. I didn’t re-read the whole file; I re-read the target: the consolidated cash flow statement in Exhibit 99.1, line by line, since that’s exactly where the falsifier broke.

What I re-derive: (1) the flow falsifier, with a diagnosis of why it was poorly chosen from the start; (2) the FCF bridge, recalculated to the second decimal from the cash flow statement; (3) all five valuation models and Monte Carlo, at a price of … instead of …; (4) a new finding from the FY2025 20-F that the previous report had only half of — the change-of-control clauses tied to the CEO’s person; (5) two factual corrections (the CEO’s age, the web brief’s figures).

The price move isn’t the thesis. … → 159.05 in the delta package, … at the 09/11/2026 close: … in 21 days, with no corporate news between the two dates. I build nothing on it; I treat it strictly as what it is — the denominator of all the margins of safety below, and the mechanical reason each MOS in this report is 5-6 points smaller than three weeks ago.

Executive summary

The thesis, unchanged in substance. Danaos is a tonnage lessor: it owns 76 containerships (485,749 TEU) and 11 Capesize bulkers, which it leases to major lines for 5-10 years. The model is plain — you buy the ship, fix it on charter, collect the freight, pay the crew and the interest, the rest belongs to the shareholder. The gap that makes the case interesting is the same as in August, just with a larger denominator: operating flow over the last 12 months is … on a market cap of …; the balance sheet has book net debt of 224.4 mil. and practically zero once liquid securities are included; the contracted backlog is … (Exhibit 99.1) and stretches to 2038; and the market pays 0.715× book value for a fleet independent appraisers hired by the company’s own lenders say has 79 of 86 vessels worth more than the balance-sheet figure.

What triggered the redo and what I found. The falsifier “if FCF yield falls below 4%, the first pillar of the thesis fails” couldn’t be evaluated because yfinance publishes only one of four quarters for DAC on the Free Cash Flow line (Q2 2026: … mil.) and doesn’t publish marketCap at all — so neither numerator nor denominator. But the more serious defect isn’t a data one, it’s a choice one: the fcf_yield metric is computed as CFO − capex, exactly the … figure the August report had explicitly rejected as unusable, since it mixes the existing fleet’s flow with yard installments for 33 undelivered vessels. A 4% threshold on that metric would measure the pace of the construction program, not the company’s ability to pay the shareholder: at 196.7 mil. and a … market cap, the yield is already 6.78% — closer to the breach threshold than to the thesis, even though nothing has deteriorated. The falsifier was set to trigger on an acceleration of investment, not a deterioration of the business. I’ve replaced it (the valuation chapter explains with what and why).

The estimated value, re-derived at … Five triangulated models give a range from …/share (own 5-year multiples) to … (DCF on … owner earnings), with a median at …, i.e. a median margin of safety of …. Monte Carlo over 20,000 scenarios, with the same central assumptions, gives a median intrinsic value of …, P10 at … and P90 at …; the probability the stock is undervalued is …%, and the probability of a margin above 30% is …%. Versus August 21, the dollar values are practically identical (models’ median 211.96 → 211.95; MC … → …) — the entire shift in margins comes from price, not fundamentals. This is exactly what should happen in a refresh with no new filings, and it’s the best check that the model wasn’t quietly rewritten.

Verdict: BUY, with a position sized for cycle and governance risk, not conviction — but with a visibly thinner margin than three weeks ago. The left tail has moved: P5 went from … to …, and P10 from … to …. In other words, in 8% of scenarios the stock is now overvalued versus my model, up from 5% in August. The … price is already above the … target Freedom Broker raised in August 2026 upon upgrading from Hold to Buy (brief, StreetInsider) and only 3.2% below yfinance’s average analyst target (164.50, on two opinions). You’re no longer buying at a discount everyone sees; you’re buying at a discount two analysts and the balance sheet see.

Suggested sizing: a normal position, no overweight. The argument remains that today’s price doesn’t require freight rates to stay where they are — at 500 mil. owner earnings, the market cap is recovered in 5.8 years (versus 5.5 in August), and the bear DCF, with a 40% flow cut, gives …, not …. The main counter-argument hasn’t changed and isn’t the cycle: it’s the ~6.0%-of-revenue rent flowing to CEO-family entities, plus the fact, newly documented below, that Coustas leaving office triggers full acceleration of the secured bank facilities and an obligation to offer to repurchase all senior notes — so succession isn’t just a governance risk, it’s a financing event.