TROX — Tronox Holdings plc · 2026-09-06 · Verdict: PASS · Conviction 3

Price $4.81 (screen row, universe_under2b.csv built 2026-09-06; no live quote available) · Mkt cap $768M · EV ~$3.80B · EV/EBIT n/m (EBIT negative CY2025 and TTM) · FCF yield -36.6% (CY2025) · Net debt $3,036M · ADV $13.1M Sources read: 10-K filed 2026-02-20 (Items 1, 1A, 7), 10-Q filed 2026-08-06 (Item 2), DEF 14A filed 2026-03-19, 8-Ks dated 2026-04-28, 2026-04-29, 2026-05-07, 2026-07-29, 2026-08-06, Form 4s (12m). No transcript or prepared remarks were filed as an exhibit, so Section 5 uses written CEO remarks, not Q&A.

1. What the business actually does

Tronox is the world's largest vertically integrated producer of titanium dioxide pigment. It mines titanium-bearing mineral sands in Australia and South Africa, upgrades them into slag and synthetic rutile, and feeds seven pigment plants in the US, Australia, Brazil, the UK, France and Saudi Arabia, serving roughly 1,200 customers in about 120 countries (10-K 2025, Item 1). In 2025 TiO2 was $2,298M of revenue, zircon $274M and other products $326M (10-K 2025, Item 7). Zircon, pig iron and the rare-earth-bearing mineral monazite fall out of the same mining, the basis of a rare earths ambition still at feasibility-study stage.

2. Why it looks mispriced, and why it is not

The screen sees negative EBIT and a negative 36.6% FCF yield because GAAP absorbed $232M of Botlek and Fuzhou closure charges in 2025 (10-K 2025, Item 7) and a $103M state deferred tax valuation allowance in Q2 2026 (10-Q Q2 2026). Underneath, momentum genuinely inflected: Q2 2026 revenue rose 19% to $868M, TiO2 volumes were the highest since Q2 2022, TiO2 and zircon prices each rose 5% sequentially, and Q3 Adjusted EBITDA is guided to $95-$115M against $73M (Q2 2026 earnings release, 2026-08-05). Forced selling is plausible too, with Moody's at B3 and S&P at CCC+ after 2026 downgrades (10-Q Q2 2026).

But the equity is not priced as though the market missed this. Against a $768M cap sits $3,036M of net debt, so buying the stock means paying about $3.80B of enterprise value. The accounting distortion is real. The mispricing is not.

3. Unit economics

The decisive fact is that the volume recovery is not reaching owners. Q2 2026 revenue rose 19% year on year while Adjusted EBITDA fell 22%, from $93M to $73M, and margin fell from 12.7% to 8.4% (Q2 2026 earnings release). Gross margin was 6.3% versus 10.8%, hurt by 2 points of price and mix, 3 points of production, idle-facility and freight cost, and 3 points of rand and Australian dollar currency (10-Q Q2 2026). First-half Adjusted EBITDA was $135M against $205M, on 11% higher revenue. Zircon volumes rose 61% in Q2 while zircon prices fell 18%, which is what absorbing rising Chinese supply looks like (10-K 2025, Item 1A).

Trailing twelve month Adjusted EBITDA is $266M, against $336M in CY2025 and $564M in CY2024, and TTM interest expense of $211M consumes 79% of it (10-Q Q2 2026). The cost programme delivered over $90M of annualised savings by end-2025 and tracks to the higher end of $125-$175M by end-2026 (10-K 2025, Item 1; Q2 2026 release), and earnings still fell, which shows how much input inflation those savings absorb. Pricing power exists mainly where trade defence does: the EU, Brazil, Saudi Arabia and India imposed anti-dumping duties on Chinese TiO2, but the Indian duties are stayed by a state court and the company concedes it cannot say whether they will be reinstated (10-K 2025, Item 1A).

4. Balance sheet and capital allocation

Total debt is $3.2B, net debt $3,036M, net leverage 11.4x, up from 9.0x at year end, and liquidity is $527M, down from $674M after the Emirates Revolver lapsed (10-Q Q2 2026). There are no financial covenants on the term loans or bonds, one springing 4.75x first-lien covenant on the revolver, and no significant maturity until 2029 (10-K 2025, Item 7). Reported net debt excludes a $50M inventory financing and a $75M June 2026 sale-and-leaseback booked as a financing obligation, and the receivables securitisation limit was raised twice in 2026 to $275M: three liquidity levers in six months.

No shares were repurchased in 2025 or first-half 2026 despite a $300M authorisation running to February 2027, and share count rose from 158.6M to 159.7M. The dividend, cut 60% in Q3 2025, is $0.05 a quarter, about $32M a year, still paid while first-half free cash flow was negative $75M (8-K 2026-07-29; Q2 2026 release). Cristal owns 24% with two board nominations (10-K 2025, Item 1A). Insiders own 2%, and over twelve months made zero open-market purchases against 162,589 shares sold at about $6.59 on 2026-03-06, including 101,600 by CEO John Romano (Form 4, 2026-03-09). Pay is at least scored honestly: incentive payouts were below target in 2022 through 2025 and performance LTIP vesting in 2025 and 2026 was forfeited below threshold (DEF 14A 2026-03-19).

5. Management: said versus did

Guidance was met at the number: Q2 was guided to $65-$85M in May and delivered $73M (Q1 2026 release, 2026-05-06). The cash promise was softer. Management expected Q2 free cash flow to be positive and "largely offsetting Q1 cash use"; Q1 used $135M and Q2 generated $60M, so the half still ran negative $75M. That $60M came from inventory, down about $120M to its lowest since June 2024. The lever is largely spent, and management is restarting a Namakwa furnace and the West Mine to rebuild inventory, which consumes working capital. Full-year "meaningful positive free cash flow" therefore rests on a large Q4 after a guided-neutral Q3.

6. Valuation

Bear, 35%: Chinese oversupply persists, Indian duties lapse, input costs stay high, 2028 Adjusted EBITDA near $300M. Interest of $210M plus capex of $250M exceeds it and net debt drifts to $3.2B. At 6.5x, EV $2.0B, below net debt: equity is option value, $0 to $1. Base, 40%: cost programme lands at $175M, pricing sticks, 2028 Adjusted EBITDA $450M, net debt flat at $3.0B. At 7.5x, EV $3.4B, equity $375M, or $2.35. Bull, 25%: a real cycle turn, duties hold, zircon stays tight, 2028 Adjusted EBITDA $600M (above CY2024), net debt $2.8B. At 8x, EV $4.8B, equity $2.0B, or $12.50. Probability weighted, about $4.24 against $4.81. Reverse DCF: at $4.81 the market pays roughly $3.80B of enterprise value, which at a 7x mid-cycle multiple requires about $543M of sustainable Adjusted EBITDA, essentially the CY2024 level and more than double the $266M earned in the last twelve months.

7. Catalysts and timeline

Q3 results in early November against the $95-$115M guide. FY2026 results in February 2027, the real test of full-year free cash flow and the cost programme. Reinstatement or abandonment of the Indian duties. The rare earths cracking and leaching feasibility study, due by Q3 2027 (Q2 2026 release). The 2029 refinancing window opening in 2027, which matters because the 2030 secured notes spring 91 days early if more than $250M of the 2029 notes remain outstanding (10-K 2025, Item 7).

8. Risks to this PASS and pre-registered kill criteria

The risk of passing is that this is a levered call option on a commodity cycle: if TiO2 prices turn hard the equity multiplies, and a funded rare earths deal would be a second re-rating path. I would revisit on any of: 1. Two consecutive quarters of Adjusted EBITDA above $115M with margin above 12%. 2. Reported net debt below $2.75B with inventory flat or higher than the prior quarter, proving deleveraging came from earnings rather than working capital. 3. The 2029 notes refinanced or extended before end-2027 at a coupon at or below 9.125%. 4. Any open-market insider purchase (Form 4 code P) by a named executive officer, or net insider buying above $1M over a rolling twelve months.

9. Verdict and summary

PASS, conviction 3. Tronox is a genuinely inflecting business attached to a capital structure that has already claimed the upside. Volumes are the highest since 2022 and prices rose 5% sequentially for both products, but Q2 revenue grew 19% year on year while Adjusted EBITDA fell 22% to $73M, so the recovery is going to price, mix, freight, currency and input costs rather than to owners. At $4.81 the equity is a $768M sliver on top of $3,036M of net debt, 11.4x trailing leverage and $211M of annual interest against $266M of trailing Adjusted EBITDA, and reported net debt still excludes a $50M inventory financing and a $75M sale-and-leaseback. Paying $3.80B of enterprise value needs about $543M of sustainable EBITDA at 7x, roughly the 2024 level and more than twice what the company earns now, so you are not paid to take the cycle risk. The details agree: three liquidity levers in six months, ratings at B3 and CCC+, a dividend paid into negative first-half free cash flow, no buybacks against a live authorisation, and insiders who sold 162,589 shares at $6.59 in March and bought none. Come back if EBITDA clears $115M twice at a 12% margin, or if net debt falls below $2.75B without emptying the warehouse.

Research for discussion, not investment advice. Positions and sizing are the reader's decision.