A levered, structurally-loss-making graphite-electrode pure-play whose old take-or-pay earnings are gone, now priced as a distressed call option on a 2026 electrode-price recovery that has to clear a 2029 debt wall — own the bonds' problem, not the equity, until pricing turns or the balance sheet is fixed.
The verdict
A levered, structurally-loss-making graphite-electrode pure-play whose old take-or-pay earnings are gone, now priced as a distressed call option on a 2026 electrode-price recovery that has to clear a 2029 debt wall — own the bonds' problem, not the equity, until pricing turns or the balance sheet is fixed.
Primary sources
SEC filings
Source documents — open to read in full
GrafTech (founded 1886, incorporated Delaware, HQ Brooklyn Heights OH) is a near-pure-play manufacturer of ultra-high-power (UHP) graphite electrodes — the industrial consumable that conducts electricity to melt scrap in electric arc furnace (EAF) steelmaking. One reportable segment: Industrial Materials. Electrodes are "the only known commercially available product" with the conductivity + heat tolerance EAF steel requires, yet represent <2% of a typical EAF's total steel-production cost — an essential, low-share-of-wallet consumable. ~96% of 2025 electrode volume went to EAF steel producers; the rest into ladle furnaces, TiO₂, silicon/ferro-alloys.
The defining structural fact: GrafTech is substantially vertically integrated into petroleum needle coke via its Seadrift plant (Port Lavaca, TX) — the key raw material. Seadrift has nameplate capacity of ~140k MT of calcined needle coke, ~one-fifth of global (ex-China) production capacity. This is the single most differentiated asset in the company.
Contract structure — the heart of the story. Historically GrafTech sold the bulk of volume under 3-to-5-year take-or-pay LTAs signed at the 2018 price peak (~$9,600/MT). The substantial majority of those LTAs have now expired, shifting the mix to short-term (annual/semi-annual/quarterly) agreements and spot. The book is built mostly in Q4 each year; there is a lag between price negotiation and revenue recognition. There is no widely accepted graphite-electrode reference price — pricing is cyclical and tracks a spread over needle coke (historical avg spread ~$4,000/MT over 2006–2025, inflation-adjusted; recent spreads much narrower).
Plain terms: GrafTech makes the carbon rods every mini-mill burns through, mines its own key ingredient, and used to lock in fat multi-year prices that have now rolled off into a brutal spot market.
Mapped upstream → company → end customer, naming every link:
Single-source dependency: Seadrift is both the company's biggest strategic moat and a concentration risk — Seadrift and Calais both sit <50 ft above sea level (climate/flood exposure flagged in risk factors).
The honest read: the Seadrift moat is durable and underappreciated, but it protects cost, not price — and the equity is dying on price.
One reportable segment (Industrial Materials) — no product-level segment P&L to disaggregate; our figures is empty. Geography (net sales mix, 2025): ~94% EMEA + Americas, ~6% APAC; ~59% of sales outside the US (vs 68% in 2024, 67% in 2023 — i.e. a deliberate shift toward the US, the strongest-priced region). Management estimates the higher US mix added ~$135/MT to 2025 weighted-average selling price. Volume trend: sales volume 109.2k MT in 2025 (+6% YoY), production 112.3k MT (+15%) — volume is accelerating while price decelerates, the central tension of this name.
The most important slide in the book: volume up, loss wider.
| Metric (Q1) | Q1 2026 | Q1 2025 | Δ |
|---|---|---|---|
| Net sales | $125.1M | $111.8M | +12% |
| Sales volume | 28.1k MT | 24.7k MT | +14% |
| Cost of goods sold | — | — | +22% YoY |
| Net loss | $(43.3)M | $(39.4)M | worse |
| Loss per share | $(1.66) | $(1.52) | worse |
| Adjusted EBITDA | $(13.5)M | $(3.7)M | sharply worse |
| EBITDA | $(3.6)M | $(4.9)M | |
| Free cash flowCash left after paying to run and maintain the business. Unlike profit, it is hard to flatter with accounting choices. | $(27.1)M | $(42.5)M | improved (working-capital timing) |
| Cash & equivalents | $120.2M | — | down from $138.4M (YE25) |
| Total stockholders' deficit | $(304.4)M | — | deepened from $(259.6)M (YE25) |
| Capacity utilization | 65% | 63% |
The damning line: COGS rose +22% on +12% revenue, so a 14%-higher-volume quarter produced a wider loss and adjusted EBITDA fell from $(3.7)M to $(13.5)M. The volume-share-gain strategy is working on volume and failing on profit. Market reaction: stock fell ~8.4% pre-market on the print; reported a wider-than-expected loss (one outlet: $(2.05) "EPS" vs $(1.42) est. — likely a non-GAAP basis; GAAP LPS was $(1.66)) while beating on revenue.
Full-year 2025 (10-K): Net sales $504.1M (−6% YoY; 2024 $538.8M, 2023 $620.5M); net loss $(219.8)M (2024 $(131.2)M, 2023 $(255.3)M); LPS $(8.45); Adjusted EBITDA $(9.1)M (2024 +$1.6M — i.e. went negative); FCF $(120.5)M; net cash used in operations $(81.6)M. Weighted-average realized price ~$4,100/MT (−13% YoY); spot ~$4,100/MT. Balance-sheet flags below (Lens 10).
Guidance/outlook: 2026 sales volume +5–10% (Q1 ~+10%), ~65% of 2026 book committed, Capital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs. ~$35M, cash COGS/MT down low-single-digit %; management concedes pricing remains "unsustainably low" into 2026.
No transcripts on the shelf (transcripts/ empty) — sourced from web call coverage. Tone arc across the last several calls: defiant-operational, not triumphant. Consistent recurring phrases: "compelling customer value proposition," "gaining market share," "shifting mix to the US," "disciplined approach of foregoing volume where margins are unacceptable," "cumulative cash-COGS/MT down 31% since end-2023." What they keep flagging and won't stop saying: pricing is "unsustainably low," competitor behavior "aggressive … arguably irrational". New on the Q1-2026 call: announced a $600–$1,200/MT price increase (effective 2026-03-26) and warned of rising decant-oil/needle-coke cost in H2 plus Middle East logistics/energy volatility. Net: management has pivoted from "wait for the cycle" to "we are actively trying to force price up" — an admission that volume alone won't fix it.
| Company | Ticker | Mkt cap | EV | EV/EBITDA | P/E | Note |
|---|---|---|---|---|---|---|
| GrafTech | EAF | ~$0.20–0.24B | ~$1.15–1.31B | NM (negative; TTM EBITDA ~$(27)M) | NM (loss) | Net debt ~$1.0B is ~83% of EV |
| HEG Ltd | HEG (NSE) | ~$1.03B | n/a | ~10x on core electrode biz (analyst SOTP) | n/a | Profitable — Q4 FY25 EBITDA margin 19.5%; owns 8.23% of GrafTech (Oct 2024) |
| Graphite India | GRAPHITE (NSE) | n/a | n/a | n/a | n/a | Indian peer |
| Resonac (ex-Showa Denko) | 4004 (TYO) | n/a | n/a | n/a | n/a | Closed Omuta JP plant late-2025; shifting to Malaysia JV |
| Tokai Carbon | 5301 (TYO) | n/a | n/a | n/a | n/a | JP peer, US presence |
Conflict flagged (do not silently resolve): some 2026 web screens still cite a GrafTech "median analyst target of $1.15" and "price $1.28" with an EV/EBITDA of −69.67. Those price levels are pre-reverse-split (1-for-10, Aug-2025); the company now trades ~$6–9. The $1.15 target ≈ ~$11.50 post-split — but I will not treat the screen-scraped $1.15 as a current target because the basis is ambiguous; the reliable, internally-consistent facts are: price ~$7.82–8.87 (Jun 9–10 2026), 52-wk range $4.92–$20.32, mkt cap ~$0.20B, and 0 Buy / 5 Hold / 0 Sell with consensus FY2026 EPS ~$(5.19). Recent rating actions: BMO raised target +$2 (May 2026); JPMorgan and RBC lowered targets (Feb 2026).
The single most useful comp line: HEG, a direct electrode peer, is profitable (19.5% EBITDA margin) and worth ~5x GrafTech's market cap on similar revenue — GrafTech's distress is company/balance-sheet specific, not purely industry-wide.
Pattern (mostly ``):
What the tape actually reacts to for EAF: (1) electrode pricing / spread direction, (2) LTA/contract structure news, (3) balance-sheet/liquidity events (refi, reverse split, equity dilution risk), (4) any China/needle-coke/battery-anode optionality headline. Earnings beats on volume don't help; the market trades the price recovery and solvency questions.
Forensic posture — this balance sheet is the whole bear case.
Regulatory findings (required sub-section).
Bottom-up from FY2025 actuals + 2026 guidance. Every line labeled; outputs ``. No our model create per --watchlist rules.
Anchors: FY25 net sales $504.1M, volume 109.2k MT, realized price ~$4,100/MT, adj. EBITDA $(9.1)M, interest $104M, D&A ~$55–62M, shares ~26.0M, capex ~$35M (2026 guide).
Revenue drivers:
Base case — modest volume (+7%), partial price capture (~+$400/MT realized net of mix, i.e. ~$4,500/MT), cash-COGS/MT down low-single-digit, decant-oil cost headwind in H2:
Bull case — full $600–$1,200/MT capture + 2026 EAF recovery (US/EU restocking), volume +10%, cost-out holds:
Bear case — price increase fails, China oversupply persists, decant-oil cost rises:
The number that matters more than EPS: cash runway vs the 2029 wall. Liquidity $328.7M (Q1-26). At a ~$(27)M/qtr FCF burn (Q1-26) that's ~3 years of RunwayHow long the cash lasts at the current rate of spending. It shortens the moment spending rises, which is why a figure taken from a quiet quarter flatters. — almost exactly to the 2029 maturities. If EBITDA inflects positive in 2026 the burn slows and the RunwayHow long the cash lasts at the current rate of spending. It shortens the moment spending rises, which is why a figure taken from a quiet quarter flatters. extends comfortably to refi; if it doesn't, they hit the wall light on cash. This is a solvency-timing bet, not an earnings bet.
Bull case. (1) Seadrift is a genuinely scarce, best-in-class, ~1/5-of-ex-China needle-coke asset with embedded option value on the synthetic-graphite battery-anode market and on US/EU China-decoupling in critical minerals — arguably worth more than the whole equity in a sane market. (2) Operating leverage is enormous off a depressed base: utilization 63%→ rising, cash-COGS/MT down 31% since end-2023; a $600–$1,200/MT price increase on ~115k MT is ~$70–140M of incremental EBITDA against a sub-$250M market cap. (3) The secular EAF tailwind is real — ~170Mt of ex-China EAF capacity expected by 2030, ~+200kt electrode demand vs ~800kt today. (4) Insider buying (CEO +212% shares) and a direct competitor (HEG) accumulating 8.23%. If price turns, this is a multi-bagger off a distressed base.
Bear case (permanent-impairment risks). (1) The balance sheet can kill it before the cycle turns — negative equity, negative interest coverage, $1.0B net debt, everything due 2029; a failed price recovery + continued ~$100M/yr cash burn forces a dilutive equity raise or a debt restructuring that wipes the equity (the market already whispers "out-of-court restructuring"). (2) Structural, not cyclical, oversupply — China electrode prices ~$1,971/MT vs GrafTech's ~$4,100; Chinese + Indian capacity expansion is permanent and "arguably irrational"; the old $9,600 LTA world is never coming back. (3) GrafTech is the high-cost, high-leverage Western incumbent — HEG/Graphite India are profitable at these prices and GrafTech is not, which means the marginal-producer pain lands here first.
Pre-mortem (18 months out, thesis broke): the $600–$1,200 price increase was only partially absorbed because Chinese exports stayed cheap and EU steel stayed weak; decant-oil costs rose, eating the cost-out; FCF stayed deeply negative; liquidity fell below ~$150M; GrafTech announced an equity raise / exchange offer at a distressed price; the stock halved again.
Are multiples too high? On current earnings, infinitely (negative EBITDA). On normalized (bull-case ~$150M EBITDA), EV ~$1.2B = ~8x — cheap if you believe normalization. The entire debate is P(normalization before the balance sheet breaks).
Contrarian view (what the market refuses to see): the market is pricing EAF as a slow-bleed equity zero, but it is mis-weighting two binary upside catalysts the consensus won't underwrite — (a) a successful electrode price increase (management is forcing it, and electrodes are <2% of customer cost, so demand is inelastic if supply discipline holds), and (b) a Seadrift monetization / strategic transaction (a needle-coke asset this scarce, inside a sub-$250M-equity shell, is an obvious carve-out or take-private target for a battery-materials or China-decoupling buyer). Either could re-rate the equity violently.
Dismantling the bull case. What breaks the money-machine: electrodes are a commodity with no reference price and no pricing power in oversupply — GrafTech doesn't set price, the marginal Chinese exporter does, and that price (~$1,971/MT) is below GrafTech's cash cost. Revenue concentration: ~94% EMEA+Americas EAF steel — a single regional steel recession (Europe is already "challenged") craters the book. Why the moat is weaker than bulls think: Seadrift insulates cost, but the bull case needs price to recover, and Seadrift does nothing for that; meanwhile the battery-anode optionality has been "coming soon" for years with no GrafTech revenue to show for it — it's a story, not a P&L line. Most dangerous competitor bulls underestimate: not China — HEG, which is profitable at these prices, owns 8.23% of EAF, and could dictate terms in any restructuring or consolidation. Worst capital allocation: the Brookfield-era leverage that loaded a cyclical commodity company with $1.1B of debt at the top — the equity holders inherited a bond-shaped risk. What must hold for today's price: a successful price increase and a 2026 EAF recovery and refinanceable credit markets in 2028–29 — three things, all outside management's control. If growth disappoints 20–30%: GrafTech doesn't have an earnings problem, it has a solvency problem, and a 20–30% revenue miss versus the recovery case means an equity raise at a distressed price — permanent dilution, not just a down year. Single scenario that permanently impairs: China holds electrode prices below Western cash cost through 2028 → GrafTech burns to the 2029 wall → debt-for-equity restructuring → equity ~zero. Plausibility: uncomfortably non-trivial — the credit market is already modeling it.
| Industry | Critical Materials |
| Size | Public Company |
Where GrafTech International sits against the other names we cover on this beat. Each line is that company’s verdict, not a summary of it.
The bear case fired — but through a door the prior dossier wasn't watching, and the stock overshot it.
Cash $1.4B
The bear case got tested on its own terms and lost
Cash $7.3B
The June thesis inverted on the arithmetic, not the story
Cash $568M
A fully-funded, government-backed option on the first US commercial-scale lithium mine
A single-asset, equity-accounted 44.8% call option on Cauchari-Olaroz