Teekay Tankers Ltd. TNK

Deep-value report: TNK — Teekay Tankers Ltd. (REDO / REFRESH regime)

Generated: 2026-09-11 · Price: … · Market cap: …B · Shares: 34.70M outstanding (30.1M Class A + 4.6M Class B), 34.91M diluted average Q2 2026 · Sector: oil shipping (crude & product tankers) · Exchange: NYSE · Currency: reporting USD / price USD (no conversion)

Regime: REFRESH over the reference thesis from 2026-08-03 (rapoarte/deep/2026-08-03-deep-TNK.md), triggered by the deep_delta.py mechanical triage on the pb falsifier (1.529 > threshold 1.5). Delta package: rapoarte/deep/delta-TNK-20260909.md.

Primary sources: FY2025 20-F (filed 2026-03-13) and FY2022/FY2023 20-Fs for the cycle series; Q1 2026 6-K (2026-05-15) and Q2 2026 6-K (2026-07-31, full financial statements) + Q2 2026 earnings release (2026-07-29) — all in SEC-Filings\TNK\; rapoarte/deep/data-pack-TNK-20260911.md (EDGAR companyfacts + yfinance); research brief rapoarte/deep/research-TNK-20260911.md; yfinance historical prices (09/11/2026); tracker Pregatire_investitii_21.xlsx via query_tracker.py.

What I inherit without re-deriving (no new SEC filing has appeared since the thesis date — the last one is the 6-K from 07/31/2026, which the August 3 report already used): the business description and the conclusion of no structural moat; the ownership and governance structure (Teekay Corp 30.7% economic / 54.8% of votes, "controlled company" status, dual CEO/CFO hat); the 2022-2025 deleveraging history; the HIGH earnings quality verdict; Kenneth Hvid's Thorndike profile.

What I fully re-derive: the FCF bridge and the owner earnings base (this is where the reference thesis's material error lies), all five valuation models, the 5-year historical multiples, the Monte Carlo simulation, and the verdict. Plus a point-by-point re-read of the 12/31/2025 → 06/30/2026 balance sheet, which didn't exist as an exercise in the old report (that one compared annual closings).

---

Executive summary (1 page: thesis, estimated value, verdict)

The falsifier that triggered the redo broke on a corrupted figure, but its substance is real. The pb > 1.5 threshold was hit on the book value published by yfinance, …/share. This is exactly the equity as of March 31, 2026 (…M / 34.644M shares = …) — a quarter behind. At June 30, 2026 real equity is …M, and shares outstanding 34.70M, so book value is …/share and the real P/B is 1.435x, below the 1.5 threshold. Literally, the falsifier didn't trigger. Economically, however, it broke worse than the triage shows: the reference claim was "conservative models (bear DCF, book NAV) already show overvaluation of ~26%," and on today's book NAV overvaluation has grown from … (August 3) to … — the price rose … in five weeks, book value only ….

The major finding of this redo, however, isn't the falsifier, but a construction error in the reference thesis's valuation base. The August 3 report used a normalized FCF of …M/year, obtained as the average of "CFO minus capex" for 2022-2025. That figure overstates earning power for two independent reasons, both verifiable line by line in filings: (1) it contains no fleet-replacement cost at all — in 2022 and 2023 TNK spent practically zero on vessels (it was paying down debt), and the average elevates to permanence a capex that cannot be zero for a company that itself writes, in the July 2026 6-K, that "approximately 65% of the fleet is 15 years old or older"; (2) it double-counts cash — the flow includes interest income (…M in 2025, …M in H1 2026 alone), and then the model separately adds the net cash position to enterprise value. Correcting both, normalized owner earnings fall from …M to …M/year (range …-175M). The difference is larger than the entire price move.

Mid-cycle earning power, built from the bottom up. Today's operating fleet (after selling the VLCC in July) has 35 tankers — 15 Suezmax and 20 Aframax/LR2, of which 32 owned and 3 chartered-in — i.e. ~12,425 revenue days/year. At a blended TCE of …/day (exactly the level realized in FY2025, a post-boom year with no major geopolitical shock) tanker net revenue is …M; plus …M from Australian marine services, STS and management (H1 2026 run-rate); minus the annualized H1 2026 cost base (…M vessel opex + …M chartering + …M G&A = …M) → mid-cycle EBITDA …M. Minus drydockings …M, freight taxes …M and normalized replacement capex …M/year → owner earnings …M. For comparison, annualized H1 2026 EBITDA is …M: today's market is running at 2.7x mid-cycle.

The valuation. Five models: bear DCF (TCE … r 12%) …/share → …; base DCF (TCE … r 11%) … → …; bull DCF (TCE … r 10%) … → …; Greenwald EPV on realized cycle earnings 2021-2026, cleaned of vessel-sale gains and interest income and scaled to today's fleet, … → …; NAV at market vessel values (anchored on the company's own 2026 transactions) … → …. Median …, range from … to …. Monte Carlo (20,000 scenarios over the same assumptions) gives a median of … and a probability of undervaluation of 0.7%. Historical multiples independently confirm this: TNK traded at an average of 0.77x book value in 2021-2025 and is today at 1.44x — the highest P/B in at least six years; year-end P/E averaged 3.81x in 2022-2025 and is today 5.68x, on peak earnings.

Verdict: AVOID AT THIS PRICE (downgrade from CANDIDATE WITH RESERVATIONS). This isn't a verdict about quality — quality is unchanged and above the sector average: a balance sheet with no drawn debt, …M net cash (37% of market cap), book value per share compounding at 22.3%/year since 2021, HIGH earnings quality, disciplined debt management. It's a verdict about price and normalization base: at … you pay ~8.2x mid-cycle EBITDA and ~1.44x a book value that is itself near the peak of the vessel-value cycle. The level at which the thesis becomes interesting again: …-70/share (parity with the base DCF and with book NAV), i.e. a 30… correction — exactly the kind of level the stock traded at in January 2026, eight months ago.

---