Cleveland-Cliffs Inc (NYSE: CLF) — Earnings Recovered Before the Balance Sheet

Published: 2026-09-11 · Verdict: Reduce · Entry price: $8 · Price target: $10 · Research confidence: High (90%)

Executive conclusion

Analyst Take

REDUCE at $12.015; base-case target $10; preferred entry $8. Cleveland-Cliffs has passed the first test in the July thesis: the earnings trough has ended. It has not passed the more important test for common shareholders: proving that the recovery will survive interest, maintenance and environmental capital, labor inflation, working-capital needs, furnace investment, and the next steel-price reversal while still retiring debt. Q2 2026 adjusted EBITDA increased to $286 million from $95 million in Q1, Q2 operating cash flow was $230 million, and management guided Q3 adjusted EBITDA to approximately $575 million. Average selling price rose sequentially by $76 to $1,124 per ton, while slab and other shipments fell 96% year over year after the unprofitable slab contract expired. These are operational facts consistent with a real price, mix, and cost inflection—not merely a change in non-GAAP presentation. [S2][S3][S5]

The balance-sheet evidence is less favorable. First-half operating cash flow was still negative $95 million, cash capital expenditure was $309 million, and simple free cash flow was therefore negative approximately $404 million. ABL borrowings increased by $444 million from year-end to $895 million, and total carrying debt rose from $7.25 billion to $7.70 billion. Debt did decline by about $60 million from March to June, so it would be too categorical to say no repair occurred during Q2; the correct conclusion is that cumulative balance-sheet repair has not yet been demonstrated. At June 30, liquidity was approximately $3.1 billion and the first senior-note maturity is in 2029, which substantially reduces immediate refinancing risk. Liquidity buys time, however; it does not eliminate the need for sustained free cash flow. [S2]

Valuation now assumes much of the earnings normalization. At $12.015 and 570.5 million shares, equity value is approximately $6.86 billion. Adding $7.70 billion of debt and subtracting $70 million of cash gives an attributable enterprise value near $14.49 billion. This deliberately excludes the $214 million noncontrolling interest because the company’s adjusted-EBITDA measure also removes EBITDA attributable to noncontrolling interests. At a 6–7x normalized steel multiple, the current enterprise value embeds roughly $2.1–2.4 billion of sustainable adjusted EBITDA. An illustrative 2026 estimate of $1.58 billion—reported first-half EBITDA of $381 million, management’s $575 million Q3 guide, and an analyst-estimated $625 million Q4—would leave the shares at about 9.2x enterprise value to estimated EBITDA. [S2][S3][S16]

The $10 base-case target assumes approximately $2.1 billion of normalized adjusted EBITDA, a 6x multiple, about $6.9 billion of forward net debt after partial deleveraging, and no further dilution, producing equity value near $5.7 billion or approximately $10 per share. The $8 entry price requires a wider discount for steel-price, execution, and capital-allocation risk. Those values are estimates, not reported facts, and they are especially sensitive to the sustainable EBITDA denominator. The upside case remains substantial because $2.8–3.2 billion of EBITDA could generate rapid deleveraging and a much larger residual equity value. The strongest counterargument to the rating is therefore not abstract tariff optimism; it is the combination of Q3 operating leverage, the slab-contract exit, 2027 fixed-price renewals, potential asset-sale proceeds, and lower unit costs.

Investment conviction is medium. Evidence quality is high for reported financials, debt, liquidity, shipment mix, pension status, incentive metrics, and current tariff terms. It is moderate for forward pricing and management’s contract-reset economics, and low for unquantified aluminum substitution, POSCO optionality, and unsupported assertions of durable market-share gains. The decision sequence is concrete: Q3 should deliver at least approximately $550 million of adjusted EBITDA; Q4 should not deteriorate if benchmark prices remain stable; second-half cash generation should reduce the ABL rather than merely finance receivables and capital spending; and 2027 realized contract pricing should validate a material portion of the claimed $500 million uplift. The call would improve after two consecutive positive-free-cash-flow quarters and cumulative net-debt reduction of at least $500 million. It would worsen if Q3 EBITDA falls below $500 million, Q4 weakens despite stable benchmarks, property proceeds do not reduce debt, labor disruption affects shipments, or another ordinary-liquidity equity issuance occurs.

Changes since 2026-07-11

The July baseline was right about the company’s capital intensity, leverage, peer-quality gap, and dependence on tariff-supported steel pricing. It was too cautious about the speed of the near-term earnings recovery. Q2 adjusted EBITDA of $286 million and Q3 guidance of approximately $575 million establish a much steeper operating inflection than Q1 alone suggested. Q2 free cash flow of approximately $73 million also falsifies the extreme bear proposition that Cliffs could not produce positive cash flow during the recovery. [S2][S3]

Four inherited claims required correction. First, deleveraging is not yet a demonstrated year-to-date fact: June debt was $450 million above year-end, notwithstanding a modest sequential decline in Q2. Second, the prior report incorrectly implied that adjusted EBITDA drove the 2025 annual-incentive payout. The EBITDA component earned zero because $37 million was below its $700 million threshold; safety and strategic metrics produced the 76% aggregate funding. Third, the earlier roughly 2.0x tangible-book framing is stale after the share-price increase and balance-sheet update; the current calculation is approximately 2.5x. Fourth, Middletown is no longer merely a reduced green-steel option. The company and Department of Energy now describe a $1 billion blast-furnace and infrastructure modernization supported by a $500 million federal award, although final implementation details and reimbursement conditions remain important. [S2][S4][S6][S18]

The bull thesis is partly confirmed: earnings and mix are improving. It is not confirmed on its original balance-sheet test because net debt remains far above $5 billion. The bear thesis is also partly falsified: core tariff protection remains in force, liquidity is substantial, and the company is not facing imminent insolvency. Labor terms, contract-reset realization, POSCO, aluminum-substitution volume, asset-sale proceeds, and sustained debt repayment remain open.

Stock Price Action — Five-Year Event Map

CLF’s price history is the visual record of operating and financial leverage. The September 11, 2026 close of $12.015 is approximately 64% below the March 28, 2022 five-year closing high of $33.07 and 106% above the May 30, 2025 five-year low of $5.83. It is approximately 26% below the October 20, 2025 52-week high of $16.18 and 54% above the March 20, 2026 52-week low of $7.82. Those price observations are facts from the historical series; the event attribution below is interpretation unless linked to a dated disclosure. [S16]

1. September 2021–March 2022: roughly $22.6 to $33.1. The price move coincided with extraordinary post-pandemic steel pricing and was subsequently reinforced by commodity and energy disruption after Russia’s invasion of Ukraine. Cliffs’ recent transformation into an integrated steelmaker gave the equity unusually high exposure to elevated North American steel spreads. The price is factual; assigning relative weight to demand, inventories, and geopolitical disruption is interpretation.

2. March–December 2022: roughly $33 to $16. Steel prices normalized, monetary policy tightened, and recession expectations increased. The fundamental point is clearer than any single daily catalyst: operating income fell from $4.03 billion in 2021 to $1.97 billion in 2022, demonstrating that peak-cycle profitability was not durable. [S1][S16]

3. 2023–early 2024: recovery toward approximately $20. The period included a recovery in steel pricing and Cliffs’ unsolicited pursuit of U.S. Steel. The failed transaction did not directly determine intrinsic value, but it revealed management’s continuing preference for scale and integrated capacity. The stock recovery preceded the later deterioration in margins.

4. Early 2024–May 2025: decline to $5.83. Cliffs closed the Stelco acquisition in November 2024 while its underlying margins and cash generation weakened. The filing record supports the interpretation of acquisition leverage colliding with a trough: 2024 operating loss was $763 million, followed by a $1.58 billion operating loss in 2025. [S1]

5. May–October 2025: rally from $5.83 to $16.18. A June 2025 proclamation increased the Section 232 steel tariff from 25% to 50% effective June 4. The timing supports a tariff-related rerating, although price action alone cannot prove causality. CLF’s larger move than less-levered peers was consistent with greater operating and financial leverage. [S13][S16]

6. October 2025–March 2026: retreat to $7.82. Full-year 2025 results reported only $37 million of adjusted EBITDA, negative $462 million of operating cash flow, and $594 million of interest expense. The reasonable interpretation is that policy optimism encountered still-poor realized economics and renewed concern about the capital structure. [S1]

7. March–September 2026: recovery to $12.015. Q1 adjusted EBITDA reached $95 million, Q2 reached $286 million, and the approximately $575 million Q3 guide materially changed the near-term earnings trajectory. The shares rose 27.8% from the prior report’s $9.40 reference price. The factor model nevertheless explains only 31.5% of historical return variance, so attributing the entire move to company execution would be excessive. [S3][S15][S16][S17]

Verdict: The chart supports neither a permanent-compounder narrative nor an imminent-insolvency narrative. It shows a high-volatility residual claim whose price anticipates changes in spreads and policy before debt reduction appears in reported accounts. The disconfirming evidence to a purely speculative reading is that the latest rally is accompanied by a filing-supported earnings inflection; the caution is that prior rallies also ran ahead of durable cash generation.

Business Overview

Operating system and economic model

Cleveland-Cliffs is a vertically integrated North American steel system. It mines iron ore; produces pellets and hot-briquetted iron; processes ferrous scrap; operates blast furnaces, basic-oxygen furnaces, and electric-arc furnaces; rolls and coats sheet and plate; and owns downstream stamping, tooling, tubing, and component operations. The 2025 filing described seven operating blast furnaces and four operating EAFs, excluding idled or closed equipment, with configured raw-steel capacity of approximately 20 million net tons after upstream idlings. It also reported ownership interests in five iron-ore mines with 20.1 million long tons of pellet capacity, a 1.9-million-metric-ton HBI facility, approximately 4.0 million tons of coke capacity, 1.8 million tons of metallurgical-coal capacity, and 21 FPT scrap-processing facilities. [S1]

The business is readily understandable at the economic-driver level: shipment volume multiplied by realized price, less metallics, coke, coal, energy, labor, freight, conversion, finishing, maintenance, and fixed overhead. What makes the equity difficult is not conceptual obscurity but nonlinear leverage. A meaningful portion of operating cost is fixed or slow to flex; fixed-price contracts delay spot-price transmission; product grades carry different premiums; individual facilities occupy different places on the cost curve; and debt converts a moderate change in enterprise earnings into a much larger change in residual equity value. [S1][S2]

Steelmaking is the dominant reportable segment. In 2025 it generated $17.95 billion of the company’s $18.61 billion of revenue and negative $16 million of adjusted EBITDA, while Other Businesses contributed $53 million. In Q2 2026, Steelmaking produced $5.05 billion of revenue and $268 million of adjusted EBITDA; Other Businesses contributed $18 million. The recovery is therefore a steelmaking recovery, not a diversified subsidiary concealing weakness in the core. [S1][S2]

Q2 2026 shipments were 4.025 million tons. Product mix was approximately 45% hot-rolled, 31% coated, 15% cold-rolled, 4% plate, 4% stainless and electrical, and less than 1% slab and other. Hot-rolled steel is the clearest commodity exposure. Cold-rolled and coated products embed additional processing and customer specification. Electrical and stainless grades can command niche premiums. Slab and other historically included the onerous third-party supply contract that expired in December 2025. The reported mix therefore provides a bridge between operational strategy and economics. [S2]

Customers and revenue stability

Cliffs sells primarily to direct automotive customers, infrastructure and manufacturing customers, distributors and converters, and other steel producers. In 2025, Steelmaking revenue was approximately $5.05 billion from direct automotive, $5.38 billion from infrastructure and manufacturing, $5.20 billion from distributors and converters, and $2.33 billion from steel producers—roughly 28%, 30%, 29%, and 13%, respectively. In Q2 2026, automotive revenue rose 17% year over year to $1.46 billion, distributors and converters rose 14% to $1.64 billion, infrastructure and manufacturing declined 4% to $1.43 billion, and steel-producer revenue declined 12% to $526 million. [S1][S2]

Approximately 35–40% of flat-rolled shipments are sold through annual or multiyear fixed-price contracts, primarily with automotive customers. Certain contracts include input-cost surcharges. The balance is not all spot: it includes index-linked and transactional arrangements with different lags. Stelco increased consolidated exposure to current North American spot markets. Fixed-price contracts smooth immediate steel-price volatility, but they do not make revenue recurring in the subscription sense. Contracts reprice, production schedules change, service centers destock, and future vehicle programs can be sourced elsewhere.

Revenue stability is low to moderate: automotive contracts delay price transmission, but most volume remains economically exposed to spot or index conditions, contract renewals, customer schedules, and input costs. Revenue declined from $22.99 billion in 2022 to $22.00 billion in 2023, $19.19 billion in 2024, and $18.61 billion in 2025 despite the Stelco acquisition. First-half 2026 revenue recovered 6% to $10.15 billion while shipments declined 4%, showing that the early recovery was led by price and mix rather than secular unit growth. [S1][S2][S16]

Automotive qualification does create customer value. Exposed-body and structural steels must meet formability, surface, coating, strength, weldability, stamping, and defect requirements. Changing a qualified supplier during a vehicle program can require testing, tooling adjustment, and supply-chain revalidation. Cliffs also offers geographic proximity and coordinated technical service. Those features reduce the risk of mid-program displacement. They do not eliminate annual price negotiation, multisharing, vehicle-platform redesign, aluminum substitution, or competition for the next program.

Vertical integration and customer value

Backward integration provides physical supply security and visibility into iron units. Cliffs’ mines supply most internal pellet requirements; its coke operations cover a substantial portion of coke demand; FPT supplies prime and obsolete scrap; and Toledo HBI supplies cleaner iron units. This system is most valuable when seaborne ore, scrap, coke, freight, or logistics become scarce. It can also let the company optimize blends across blast furnaces and EAFs.

The same integration raises fixed and replacement capital. An integrated route requires pellets or HBI, coke, scrap, natural gas, electricity, rail, ports, environmental controls, and synchronized operation across mines, furnaces, casters, rolling lines, and coatings. Blast furnaces cannot be turned off and restarted as simply as an EAF melt shop. Mines and coke batteries continue to carry cost when internal steel output falls. Integration is therefore an insurance asset whose value must be tested against the return on the entire capital base, not assumed from physical self-sufficiency.

Cliffs’ four EAFs and HBI plant complicate the simplified description of the company as a pure blast-furnace producer. It has route optionality and can use different metallics in different assets. Nevertheless, the consolidated financial record retains the high fixed-cost and maintenance characteristics of a legacy integrated network. The company’s own idlings and configuration reductions show that not all physical capacity is economically equivalent.

Unrecognized assets and accounting limits

Economically valuable assets not fully recognized on the balance sheet include mineral reserves, permits, automotive qualifications, metallurgical recipes, operating know-how, customer engineering relationships, and integrated logistics. Those assets would be expensive and slow to recreate, but their value is conditional on profitable utilization. They cannot simply be added to book value at replacement cost because a replacement-cost asset can still earn below its cost of capital. [S1]

Conversely, balance-sheet equity contains $1.78 billion of goodwill and $1.06 billion of other intangibles at June 30, 2026. Tangible book value is therefore much lower than reported equity, and neither measure is a liquidation appraisal. The single-segment presentation also limits analysis: investors do not receive current stand-alone profitability for Stelco, mining, HBI, scrap, automotive sheet, electrical steel, individual blast furnaces, or individual EAFs. Claims that a specific acquired asset or niche is highly profitable cannot be independently reconciled to segment returns.

The company consolidates SunCoke Middletown as a variable-interest entity even though it owns no equity, because it purchases all coke and power under long-dated agreements. The VIE’s creditors do not have recourse to Cliffs’ general assets, but the arrangement illustrates how physical integration can create durable economic commitments outside conventional ownership.

Security and tax status

CLF is ordinary common equity of an Ohio corporation listed on the NYSE; it is not an ADR, master limited partnership, pass-through partnership, or K-1 security. No preferred shares were outstanding at year-end 2025. Investors receive ordinary corporate-equity tax reporting, subject to their own circumstances. [S1][S16]

Verdict: Cliffs owns strategically relevant assets and provides real customer value in automotive qualification, electrical steel, reliable delivery, and integrated raw-material supply. The disconfirming evidence to a broad franchise claim is consolidated profitability: those capabilities did not prevent negative gross margins, negative free cash flow, or substantial idlings at the trough. The best description is a high-fixed-cost steel system with narrow differentiated profit pools—not a recurring-revenue or asset-light franchise.

Industry Dynamics

Relevant market, geography, and demand

Cliffs’ economically relevant market is North American flat-rolled sheet, plate, electrical steel, and selected metallics, rather than global crude-steel production in the abstract. The United States is the dominant market, with Canada becoming material through Stelco. Automotive production, service-center inventory cycles, construction, machinery, appliances, energy infrastructure, transformers, and general manufacturing determine demand. Exports are not the central profit pool; domestic and cross-border North American prices, utilization, logistics, and trade rules are.

The addressable market is large, mature, and principally North American; utilization and spread matter more than high secular volume growth. OECD analysis expects global steel demand to grow slowly while capacity expands more rapidly. Electrical-grid investment, transformers, data centers, and selected advanced automotive grades may grow faster than aggregate demand, but conventional sheet remains linked to mature manufacturing and construction cycles. [S1][S10]

The global capacity backdrop defines the competitive boundary even when tariffs limit direct imports. The OECD estimated global excess steelmaking capacity around 640 million metric tonnes in 2025 and projected approximately 745 million tonnes by 2028. It identified up to 138.8 million tonnes of gross additions planned for 2026–2028, against much slower demand growth, and projected global utilization could fall from approximately 76% in 2025 to 74% or less by 2028. China exported approximately 131 million tonnes in 2025. These estimates are scenarios rather than guarantees—projects can be delayed and legacy capacity can close—but they establish persistent pressure for exports and state support. [S10]

Market structure and profit pools

North American steel combines a concentrated group of major domestic producers with intense product-level competition and a globally fragmented supply base. Nucor and Steel Dynamics are the most useful public operating peers because they compete in North American sheet and use predominantly EAF production. Nippon-owned U.S. Steel is a direct integrated and automotive competitor, although current public comparability changed after acquisition. ArcelorMittal and Gerdau provide broader global and regional context. Reliance is useful for downstream processing economics but is not a primary steelmaking peer.

Profit pools differ by route and product. EAF producers can generally vary melt schedules more flexibly, use scrap and alternative metallics, and carry less blast-furnace and coke infrastructure. Integrated producers retain capability in demanding exposed automotive grades and electrical steels and can benefit from captive ore when scrap or metallics tighten. Fabrication and downstream processing can earn steadier margins because customers pay for inventory availability, cutting, coating, forming, logistics, and reduced complexity. Specialty electrical grades can produce attractive premiums but remain too small to determine Cliffs’ consolidated return.

The industry has high physical barriers but weaker economic barriers. A new mill requires billions of dollars, permits, land, reliable power and water, rail access, skilled labor, customer qualification, and years of construction. Those obstacles deter casual entry. Yet attractive peak spreads, government incentives, and the lower operating cost of modern EAF facilities continue to attract investment. State-supported foreign capacity can also enter without the return thresholds expected of private U.S. capital. High physical barriers therefore do not ensure scarcity rents for legacy producers.

Industry barriers are high operationally but porous economically: permits, capital, logistics, and qualification constrain entry, while subsidized foreign capacity and modern domestic EAF investment continue to add or replace supply. The industry can generate extraordinary peak-cycle returns but has difficulty sustaining them because elevated spreads invite capacity, inventory, substitution, and customer resistance. [S10][S11][S12]

Capital-cycle test

The capital-cycle evidence is unfavorable to the idea that current protected spreads can be capitalized indefinitely. New North American capacity is being built with modern technology and directed at attractive flat-rolled and downstream markets. Gross capacity announcements are not the same as net supply: blast-furnace closures, delays, lower utilization, and project cancellations may offset additions. But even replacement rather than net growth can lower the cost curve and improve competitors’ quality capability.

Cliffs’ own actions are evidence. Configured raw-steel capacity declined from approximately 23 million tons in 2024 to 20 million in 2025 after the Dearborn upstream idling. Minorca was idled and Hibbing partially idled to consume excess pellet inventory. Steelton and Weirton generated restructuring costs, and several facilities operated below potential. A company removing marginal legacy tons while better-capitalized peers invest is on the defensive side of the capital cycle. [S1]

Current peer results show that the tariff-supported price environment does not accrue equally. Nucor reported Q2 2026 EBITDA of $2.02 billion, net income attributable to shareholders of $1.16 billion, $2.69 billion of cash and short-term investments, and an undrawn $2.25 billion revolver. Its quarter benefited from approximately $130 million of raw-material refunds and a $61 million noncash Helion valuation gain, so the headline EBITDA should not be treated as a perfectly clean run rate. Steel Dynamics reported $721 million of steel-operations income, $921 million of consolidated adjusted EBITDA, and a management-calculated three-year after-tax ROIC of 13%, while funding its aluminum ramp and repurchasing $200 million of shares. Cliffs reported $286 million of adjusted EBITDA and remained loss-making after depreciation and interest. [S3][S11][S12]

Competition is becoming more demanding, not less: tariffs reduce import pressure, but modern domestic capacity, a recapitalized U.S. Steel, and better-capitalized EAF peers continue to challenge legacy integrated assets. The disconfirming evidence is that closures can offset gross additions and automotive qualification slows immediate share transfer. The correct conclusion is rising capability competition, not a mechanically forecast net glut.

Foreign production and tariff regime

Foreign low-cost production remains a structural threat because steel is transportable and global excess capacity creates continuing pressure to export directly or through derivative products. “Low cost” is not merely low wages; it can reflect energy, ore, scale, currency, subsidies, environmental standards, financing, and utilization. The OECD links non-market support and excess capacity to weak profitability and trade distortion. [S10]

The June 2025 U.S. proclamation increased the additional Section 232 duty on covered steel articles from 25% to 50%, with specified exceptions. Subsequent 2026 proclamations changed the scope and rate treatment for certain derivative categories and trading-partner circumstances. Core steel articles remain strongly protected, but a blanket statement that every steel-containing import is uniformly taxed at 50% is false. Policy now distinguishes core metal, specified derivatives, metal content, origin, and partner arrangements. [S13][S14]

Trade protection raises the landed price of imports and can support domestic utilization. It does not eliminate domestic competition, guarantee end demand, prevent substitution, or stop circumvention. It benefits every qualifying domestic producer and is therefore not a Cliffs-owned moat. The 2026 adjustments also demonstrate that the regime can change at the category level without wholesale repeal. Valuation should treat tariffs as a material current earnings support whose durability and precise application require monitoring—not as perpetual property.

Pricing and inputs

Hot-rolled coil is the most visible benchmark, but it is not a complete revenue model. Cliffs’ 2025 filing said domestic HRC averaged approximately $851 per ton, up from $772 in 2024, while Cliffs’ realized steel-product price fell to $1,005 from $1,081 because fixed-price contracts, product premiums, lagged pricing, Stelco, slabs, and customer mix differed from the index. Q2 2026 realized price reached $1,124. Applying a spot benchmark mechanically to all tons would therefore produce a false earnings estimate. [S1][S2]

Inputs create a second spread. Blast-furnace operations consume pellets, coke, metallurgical coal, scrap, natural gas, electricity, oxygen, alloys, and logistics. EAF operations are more directly exposed to scrap, metallics, and power. Cliffs’ vertical integration reduces some exposures but does not remove opportunity cost or operating expense. The Q2 filing disclosed HRC derivatives covering 583,450 tons through September 2027—small relative to annual shipments—so the equity remains materially exposed to steel-price realization. [S2]

Regulation and bargaining power

Environmental permits, mining reclamation, emissions controls, workplace rules, trade remedies, and labor agreements are financially material. Regulation creates entry barriers because a greenfield operator must obtain complex approvals. It also creates incumbent liabilities and capital obligations. The same environmental infrastructure that makes an operating site scarce can require years of remediation and recurring spending.

Customer bargaining power is high in automotive and distribution. Large OEMs have scale, detailed cost information, multiyear sourcing calendars, and the ability to qualify multiple suppliers. Service centers can reduce inventories rapidly when prices fall. Supplier bargaining power varies: captive iron ore and coke reduce dependence, but electricity, gas, rail, alloys, and specialized maintenance remain external. Organized labor has meaningful bargaining power because plant-specific skills and synchronized operations make disruption costly.

Verdict: The industry can support several profitable years under tariffs, infrastructure demand, and disciplined closures, which is the strongest disconfirming evidence to a uniformly bearish structure. The through-cycle evidence remains unfavorable: demand is mature, global excess capacity is rising, modern domestic capacity is entering, and price signals cause capital to return. Current margins should therefore be normalized conservatively rather than capitalized as a permanent policy rent.

Competitive Position

Sources of differentiation

Cliffs has three credible operating advantages. First, it is a major automotive-sheet supplier with long-standing qualification, coating, metallurgy, logistics, and stamping expertise. Second, it describes itself as the sole U.S. producer of grain-oriented electrical steel and also produces non-oriented electrical steel. Third, it owns an integrated ore, pellet, HBI, coke, coal, scrap, steelmaking, finishing, and components network. [S1]

Each advantage has an observable financial test. Automotive qualification should preserve program volume, reduce defects, support premium mix, and improve renewal economics. Electrical-steel capability should produce attractive pricing and utilization relative to commodity sheet. Vertical integration should reduce input volatility or lower delivered cost enough to support through-cycle margins. If none appears in cash margins and ROIC, the capability is strategically useful but not a consolidated moat.

Brand, customer captivity, and switching costs

Brand has little stand-alone economic value; metallurgical qualification, coating capability, defect performance, delivery reliability, and customer engineering are the relevant sources of switching friction. End consumers generally do not pay more for a vehicle or appliance because its steel came from Cliffs. Automotive awards are evidence of execution, not consumer pricing power. [S1][S15]

Automotive switching costs are material within a qualified vehicle program but reset when customers source new platforms, redesign components, or qualify alternative suppliers. A mid-program change can require testing and create defect or production risk, while a future platform can be dual-sourced or assigned elsewhere. The mechanism provides temporary customer stickiness rather than permanent captivity. [S1][S2]

Q2 automotive revenue increased 17%, and management said automotive shipments were the highest in two years. The filing independently verifies revenue improvement but does not disclose program-level market share, contribution margin, or whether gains arose from industry production, customer inventory, competitor disruption, or permanent sourcing changes. Durable share gain would require several renewal cycles of volume, price, quality, and margins—not a single quarter.

Nature of competition

Competition is primarily delivered price, product specification, quality consistency, lead time, service, and supply reliability—not advertising or consumer brand. Commodity hot-rolled tons compete mostly on price and availability. Exposed automotive grades add surface quality, formability, coating, and qualification. Electrical steel adds magnetic-performance capability and strategic domestic availability. Slab and metallics compete more directly on specification, freight, and price. [S1][S10]

Cliffs competes with Nucor and Steel Dynamics across an expanding range of flat-rolled products; with Nippon-owned U.S. Steel in integrated and automotive sheet; and with imports when duties and logistics permit. Aluminum and other materials compete in selected automotive, construction, and packaging applications. Trade protection raises the umbrella for all domestic steelmakers and does not improve Cliffs’ relative position against better-cost domestic peers.

Cost curve and flexibility

Through-cycle outcomes provide the best public cost-position evidence because plant-level costs are not disclosed. Company Financials reports Cliffs’ ROIC at approximately 33.6% in 2021, 12.7% in 2022, and 4.2% in 2023 before returns turned negative in 2024 and 2025. Gross margin compressed from 22.2% in 2021 to 11.0% in 2022, 6.2% in 2023, 0.3% in 2024, and negative 4.6% in 2025. [S1][S16]

Nucor and Steel Dynamics remained profitable through the same downturn and entered the recovery with the balance-sheet capacity to fund growth and return capital. Their diversification and product mixes prevent a perfect comparison, and Nucor’s latest quarter included favorable nonrecurring items. Even after those qualifications, the gap is too persistent and too large to explain solely by accounting or end-market mix. It supports an inference that EAF flexibility, downstream diversification, and balance-sheet strength provide superior through-cycle economics. [S11][S12]

Cliffs’ captive ore can outperform when seaborne ore or prime scrap is scarce, while integrated furnaces retain capability in certain demanding grades. Its HBI and EAF assets provide some metallics and route flexibility. Conversely, blast furnaces, coke batteries, mines, and finishing networks carry fixed costs; idling and restarting them is costly; and environmental capital is substantial. The 2025–2026 footprint actions confirm that marginal assets can become uneconomic well before the company as a whole runs out of liquidity.

GOES, NOES, and narrow niches

Grain-oriented electrical steel is the strongest narrow-moat candidate. Transformer laminations require specialized chemistry, processing, and magnetic performance. Domestic production has strategic value, and the Butler investment may expand capability. Stainless and electrical steel shipments together were 159,000 tons in Q2, approximately 4% of total volume. Even if margins are excellent, this niche cannot alone determine returns on a system carrying $7.7 billion of debt. [S2][S15]

Non-oriented electrical steel can benefit from motors and electrification, but competitors can invest and customers can qualify imports or alternative specifications. The correct analytical treatment is a differentiated growth pocket whose value must be demonstrated through volume, premiums, and cash returns, rather than a reason to assign the entire company a specialty-materials multiple.

Aluminum substitution

Management argues that advanced high-strength steel can replace aluminum in selected vehicle applications while using equipment originally intended for aluminum stamping. The mechanism is plausible: steel may lower material cost, simplify supply, or improve lifecycle economics while meeting weight and safety requirements. Management has not disclosed annual tons, model count, realized pricing, capital required, or contribution margin. [S5][S15]

Without those data, aluminum substitution is an option rather than a forecast. Evidence that would upgrade it includes named production programs, recurring shipment volume, incremental contract revenue, and margins clearly above ordinary automotive sheet. Marketing demonstrations and customer discussions are not equivalent to production economics.

Barrier-to-entry conclusion

A barrier is valuable only if the company would lose measurable economics without it. Without automotive qualification, Cliffs would lose program volume and premium mix. Without electrical-steel capability, it would lose a strategic specialty pool. Without captive iron units, input volatility and supply risk would rise. Yet the company owned all three during years of negative gross margin and cash burn. The capabilities are necessary for relevance but insufficient for excess consolidated returns.

Cliffs has customer-specific and asset-specific advantages, but no demonstrated consolidated moat protecting returns on the full invested-capital base. The strongest disconfirming evidence is the rapid 2026 margin inflection after facility rationalization and contract exit. That could indicate a permanently lower break-even point. It should be tested through a weaker price environment before being classified as structural rather than cyclical.

Verdict: Competitive position is mixed operationally and weak economically. Cliffs is strategically important and differentiated in automotive and electrical steel, but its cost flexibility, balance sheet, and through-cycle ROIC remain inferior to the leading EAF peers. The investment question is whether the reorganized footprint can earn above its cost of capital at ordinary mid-cycle prices, not whether it can produce excellent EBITDA during a protected upswing.

Growth History and Forward Opportunities

Scale without proven economic growth

Cliffs transformed from an iron-ore producer into an integrated steelmaker through AK Steel, ArcelorMittal USA, FPT, and Stelco. Revenue expanded dramatically after the 2020 acquisitions and reached $22.99 billion in 2022, but then declined for three consecutive years to $18.61 billion in 2025 despite the addition of Stelco. The company became larger and more vertically integrated; the 2024–2025 losses show that accounting scale did not guarantee economic growth. [S1][S16]

The favorable historical interpretation is that the 2020 transactions were completed near a cycle trough, positioned Cliffs for extraordinary 2021–2022 cash generation, and enabled substantial debt repayment. The unfavorable interpretation is that management later treated peak cash flow as sufficiently durable to support repurchases and another large acquisition, thereby rebuilding leverage before the 2025 trough. Both interpretations are supported by the sequence.

Product outlook and operating levers

The product outlook is a price-and-mix recovery rather than secular volume growth: hot-rolled and coated products benefit most from current pricing, automotive from contract resets, and electrical steel from grid investment, while commodity slab and marginal integrated capacity remain weaker. [S2][S5][S10]

The first lever is the expired slab contract. The roughly 1.5-million-ton annual arrangement expired in December 2025 after being unprofitable in 2024 and 2025. Q2 slab and other shipments declined to 16,000 tons, down 96%, while hot-rolled and coated shipments increased. The filing attributed approximately $200 million of year-over-year revenue improvement to favorable mix. This validates the operational direction but not the entire annual EBITDA estimate previously associated with the contract, because benchmark prices, utilization, idlings, and replacement customers changed simultaneously. [S1][S2]

The second lever is 2027 automotive contract renewal. Management estimated that resetting fixed-price contracts at conditions prevailing during the July call could add approximately $500 million of annual EBITDA. That is the largest identified 2027 bridge, but it is a management claim rather than contracted public evidence. Investors do not know the exact tonnage, starting price, surcharge mechanics, customer mix, volume trade-offs, or overlap with current pricing. [S5]

The third lever is footprint savings and utilization. Idling uneconomic assets can remove fixed losses, while redirecting melt to better downstream lines can improve mix. The potential is real because incremental volume has high contribution over a fixed network. The risk is that savings disappear into wage inflation, maintenance, energy, or underutilized remaining assets. Unit costs and cash margins over several quarters are the appropriate test.

The fourth lever is electrical steel. Transformer replacement, grid investment, motors, and defense procurement can support GOES and NOES demand. The Butler project is scheduled for 2028 completion according to management. The opportunity is strategically attractive, but electrical and stainless tons remain a small percentage of consolidated shipments, and project return depends on actual premiums and utilization. [S15]

The fifth lever is aluminum substitution. It could create share gains without greenfield steel capacity if customers use existing stamping infrastructure. Because management has not quantified production tons or margins, it remains excluded from the base case. The sixth lever is Stelco normalization: Lake Erie hot-rolled production may benefit from improved Canadian trade conditions, but downstream Canadian finishing remained weaker in management’s Q2 discussion and no stand-alone EBITDA is disclosed. [S5]

Middletown modernization

The August announcements define Middletown more clearly. The company and Department of Energy describe a $1 billion investment supported by a $500 million federal award. The scope includes a blast-furnace rebuild, raw-material handling, process controls, energy efficiency, and cogeneration, with deployment over roughly four years and targeted furnace completion in the first quarter of 2030. The company says the facility represents approximately three million tons of raw-steel capacity and about 2,300 jobs. [S6][S18]

The project may improve reliability, yield, energy recovery, and cost. It also extends economic commitment to blast-furnace production. Federal support cuts the company’s stated share of gross project cost approximately in half, but it does not guarantee utilization, steel pricing, customer offtake, maintenance savings, or schedule. The company described a framework for final negotiations and implementation, so final milestones, reimbursement timing, eligible costs, and clawbacks remain relevant.

Property sales and partnerships

Management discussed approximately $400 million of expected property-sale proceeds, with roughly $70 million received by the Q2 call and most of the remainder expected during the second half. Asset sales can accelerate debt reduction and monetize nonproductive land. They are not recurring operating cash flow, and proceeds should be judged by the resulting net-debt change after taxes, transaction costs, capital expenditure, and working capital. [S5]

POSCO remains unpriced optionality. A partnership could provide capital, validate assets, or create customer access. It could also introduce governance complexity, financing commitments, or dilution. The dialogue had not produced definitive public economics by the report date, so no base-case value is assigned.

Verdict: Earnings can grow rapidly through price realization, slab exit, fixed-contract resets, utilization, and cost actions. The disconfirming evidence to a purely cyclical view is the potential growth in electrical steel and aluminum substitution. Neither is disclosed at sufficient scale to transform the consolidated growth profile today. The near-term opportunity remains recovery and deleveraging, not dependable high-return unit expansion.

Financial Quality

Multi-year reported performance

The financial record captures an unusually wide cycle. Values below are reported GAAP figures except adjusted EBITDA and calculated free cash flow. Company Financials was reconciled to the filings, including the 2025 filing’s restated historical presentation where applicable. [S1][S2][S16]

The series shows price and fixed-cost leverage more clearly than any qualitative label. Revenue declined about 19% from 2022 to 2025, while gross profit moved from $2.52 billion to negative $860 million. Free cash flow swung from positive $1.48 billion in 2022 to negative $1.02 billion in 2025. The downside was amplified by acquisition debt and the inability to reduce fixed cost at the same speed as realized pricing.

Earnings are emerging from a severe cyclical trough, but reported and first-half returns have not yet reached a normalized mid-cycle level. Q2 and Q3 guidance establish a rapid recovery; they do not establish the duration of margins or the equity’s through-cycle earning power. [S1][S2][S3]

Q2 earnings bridge and quality

Q2 revenue was $5.226 billion, up from $4.922 billion in Q1. Consolidated net loss was $134 million, while net loss attributable to Cliffs shareholders was $145 million. Adjusted EBITDA was $286 million, compared with consolidated GAAP EBITDA of approximately $264 million. The difference between improving EBITDA and continuing net loss is primarily depreciation, interest, and smaller adjustment items attached to a large asset and debt base. [S2][S3]

Steelmaking gross margin was 2% and adjusted EBITDA margin 5%. Average selling price rose to $1,124 per ton. The year-over-year revenue bridge included roughly $240 million from price, $200 million from favorable product mix, and $110 million from other factors, offset by approximately $270 million from lower volume. This supports the claim that slab exit and automotive mix mattered, but it also shows that price—not volume growth—was central.

Adjusted EBITDA is useful for comparing operating momentum, but it is not owner earnings. It adds back interest, taxes, depreciation, idling and restructuring charges, acquisition items, and noncontrolling-interest effects. Some costs are clearly unusual; repeated idlings and restructuring across a changing footprint are economically less exceptional. Depreciation cannot be dismissed in a network requiring relines, environmental systems, mine development, rolling-line maintenance, and replacement equipment.

Q2 operating cash flow of $230 million less $157 million of cash capital expenditure produced approximately $73 million of simple free cash flow. That is an important positive inflection. It is one quarter, and first-half free cash flow remained negative $404 million. Cash conversion should be evaluated cumulatively because working capital can move sharply as steel pricing and volumes change.

ROIC and normalization

Business profitability is highly cyclical: reported ROIC fell from approximately 33.6% in 2021 to 12.7% in 2022 and 4.2% in 2023 before becoming negative in 2024 and 2025. [S1][S16]

Peak ROIC was real cash profitability, but it is not a normal estimate. A through-cycle calculation should retain acquisition consideration and the capital tied up in mines, furnaces, working capital, goodwill, and acquired customer relationships. Writing down goodwill can mechanically improve future reported ROIC without recovering the cash originally invested. Research-adjusted judgment should therefore keep the economic acquisition cost in the denominator.

A rough cash-earnings threshold illustrates the problem. Cliffs must cover approximately $1.1 billion of annual depreciation, roughly $600 million of interest at the recent run rate, $700–900 million of capital expenditure depending on relines and project timing, pension and OPEB cash contributions around $125 million, working-capital volatility, and cash taxes when losses are exhausted. Depreciation and capital spending are not identical in any one year, but the scale of both matters.

At approximately $1.5–1.6 billion of adjusted EBITDA, reported earnings improve markedly yet owner cash remains constrained after interest and capital expenditure. Around $2.1–2.4 billion, material deleveraging becomes plausible if working capital is neutral. Above $2.8 billion, financial leverage can produce rapid equity-value growth. Below approximately $1 billion, interest and maintenance needs can again consume or exceed operating cash. These are analytical ranges rather than company forecasts.

Balance sheet, liquidity, and refinancing

At June 30, 2026, carrying debt was $7.703 billion: $6.808 billion of senior notes and $895 million under the ABL facility. Cash was $70 million, producing net debt of approximately $7.63 billion. The company reported approximately $3.014 billion of ABL availability and total liquidity near $3.1 billion. The ABL matures in June 2028. Principal maturities were approximately $895 million in 2028, $1.268 billion in 2029, $750 million in 2030, and $4.86 billion thereafter. [S2]

The springing fixed-charge covenant did not apply at June 30 because availability exceeded the relevant threshold. This reduces near-term covenant risk. A borrowing-base facility, however, depends on eligible receivables, inventory, and equipment. A severe downturn can weaken operating cash and collateral availability simultaneously. The absence of a 2026–2027 bond maturity is meaningful protection, but refinancing work will occur before legal maturity.

Debt rose by approximately $450 million from year-end, driven primarily by $444 million of net ABL borrowing in the first half. It declined modestly from Q1’s reported level to June. The fair characterization is therefore: recovery stopped part of the first-quarter deterioration but had not repaired the year-to-date balance sheet. Asset-sale proceeds could accelerate repayment, but the evidence test is reported net debt after closings—not management’s gross-proceeds expectation.

Net income and operating cash flow

Net income and operating cash flow diverge because depreciation, deferred taxes, pensions, restructuring, and working capital are large; in first-half 2026, higher receivables absorbed much of the operating recovery. [S2]

First-half consolidated net loss was $363 million. Depreciation and amortization added back $521 million. Accounts receivable consumed $607 million as pricing and sales recovered; inventories released $229 million; and accounts payable and accrued expenses supplied $277 million. After other items, CFO remained negative $95 million. The receivables build is not necessarily evidence of bad collections, but it demonstrates why rising EBITDA does not instantly equal cash.

In a downturn, working capital may release cash as receivables and inventories shrink, partially cushioning earnings weakness. During a recovery, the reverse can occur. Investors should therefore separate recurring cash margin from temporary balance-sheet flows and monitor receivables, inventory, payables, and customer terms alongside EBITDA.

Accounting quality

The company uses LIFO for portions of inventory, established depreciation methods by asset class, impairment testing for goodwill and long-lived assets, and purchase accounting for Stelco. A 2025 LIFO decrement increased cost of goods sold by $57 million. The 2025 filing also corrected immaterial employment-cost accrual errors affecting earlier interim periods and years 2022–2024. The proxy disclosed recovery of a prior incentive overpayment associated with a 2023 accounting correction. [S1][S4]

Accounting is broadly normal for an integrated steel producer, neither obviously aggressive nor maximally conservative; the central analytical risk is treating recurring reinvestment and repeated restructuring as irrelevant because adjusted EBITDA excludes them. [S1][S2]

GAAP can understate near-term cash earning power when depreciation exceeds current cash capital expenditure, particularly in a cash-preservation year. It can also correctly warn that assets are being consumed. Adjusted EBITDA can help identify a cycle turn, but it overstates distributable economics if maintenance, idling, environmental, and pension costs are treated as permanently optional. Investors should use both.

Capital intensity and other obligations

The business is highly capital intensive: cash capital expenditure was $561 million in 2025 and $309 million in first-half 2026, the filing anticipated approximately $900 million over the next twelve months, and Middletown adds a multiyear company-funded commitment. [S1][S2][S6]

The $900 million twelve-month figure includes project and reline timing beyond the $700 million calendar-2026 guidance, so the two figures are not necessarily contradictory. Burns Harbor’s furnace work, Middletown, Butler, mines, environmental systems, and routine maintenance create a persistent claim on cash.

Material obligations beyond funded debt include leases, take-or-pay raw-material and logistics contracts, environmental and reclamation duties, letters of credit, surety support, and potential multiemployer-plan withdrawal exposure. [S1]

At year-end 2025, the defined-benefit pension plans had $4.346 billion of assets against $4.131 billion of obligations, a $215 million funded surplus. OPEB plans had $763 million of assets against $1.277 billion of obligations, a $514 million deficit. Expected 2026 cash contributions were $43 million for pension and $83 million for OPEB. Pension is therefore not the principal solvency risk, although OPEB and individual-plan funding still require monitoring.

Verdict: Financial quality is improving rapidly from a weak base. Liquidity is substantial, the pension headline is manageable, and Q2 cash flow was positive. The disconfirming evidence to a completed recovery is decisive: first-half free cash flow remained negative, debt was above year-end, Q2 gross margin was only 2%, and the asset base requires heavy recurring capital. Sustainable ROIC above the cost of capital remains unproven.

Capital Allocation

Acquisitions

AK Steel and ArcelorMittal USA transformed Cliffs in 2020, followed by the $775 million FPT acquisition in 2021 and Stelco in 2024. The 2020 transactions were well timed relative to the subsequent steel-price boom and generated strategic scale, automotive relationships, and finishing capability. Public reporting does not provide a clean asset-level return series that separates purchase price, assumed obligations, follow-on capital, and synergies.

Stelco’s accounting purchase consideration was $3.208 billion: $2.450 billion of cash, $343 million of Cliffs shares, and $415 million of debt consideration. It added Canadian hot-rolled production and greater spot exposure. Post-close 2024 results included $329 million of revenue and a $58 million attributable loss; current stand-alone EBITDA and maintenance capital are not disclosed. [S1]

The acquisition record created strategic scale but has not demonstrated attractive through-cycle returns; Stelco remains particularly unproven because purchase consideration is disclosed while current stand-alone cash returns are not. [S1]

FPT strengthened scrap sourcing and route flexibility, but the later sale of FPT Florida for $53 million shows that not every acquired asset remained strategically essential. Acquisitions should be judged on cash-on-cash returns across a cycle, not tonnage, revenue, or strategic descriptions alone.

Repurchases, issuance, and dilution

Cliffs repurchased 37.9 million shares during 2024 for approximately $733 million, an average near $19.34 per share. In October 2025 it issued 75 million shares at $12.69, generating approximately $951 million of net proceeds used to repay ABL borrowings. This was an underwritten public offering, not a debt-for-equity exchange. Shares outstanding increased from 493.9 million at year-end 2024 to 569.8 million at year-end 2025 and 570.5 million at June 2026. [S1][S2]

The company is not currently executing a material buyback: authorization remains available, but the 2025 public offering more than reversed the 2024 share-count reduction. [S1][S2]

The sequence—repurchase at a higher average price followed by issuance at a lower price—reduced per-share value relative to retaining capital, although the later offering improved liquidity and was defensible once the balance sheet had weakened. The lesson is not that all repurchases or issuance are wrong; it is that cycle-aware price discipline matters more for a leveraged commodity producer.

Material recent dilution came from a public capital raise, not issuance to insiders, although equity compensation remains part of executive and director pay. [S1][S4]

Free cash flow, dividends, and reinvestment

Free cash flow was negative approximately $1.02 billion in 2025 and negative $404 million in first-half 2026; management identifies debt reduction as the priority, but cumulative reported debt had not declined by June. [S1][S2][S5]

Property sales can improve the second-half debt path but are nonrecurring. Project spending can be rational if expected returns exceed the cost of capital, but the burden of proof is higher when debt is substantial. A credible allocation hierarchy would fund safety and maintenance, preserve liquidity, complete only high-return projects, and direct residual cash to debt before resuming repurchases or major acquisitions.

The common dividend remains suspended and is not covered by current cumulative free cash flow; debt repayment and required reinvestment should remain senior uses of cash. [S1][S2]

Compensation, governance, and insider activity

The 2025 annual incentive weighted adjusted EBITDA 50%, strategic initiatives 40%, and safety 10%; adjusted EBITDA missed its $700 million threshold and contributed zero to the payout. [S4]

Safety funded at 200% of target and contributed 20 percentage points; strategic initiatives funded at 140% and contributed 56 points; total annual-incentive funding was 76%. Long-term awards combined performance cash, performance shares, and RSUs. Performance awards depend on three-year relative TSR, with zero below the 25th percentile and up to 200% at the 75th percentile. CEO total compensation was approximately $19.0 million in 2025.

The scorecard contains objective thresholds, but the strategic component permits judgment and relative TSR can reward outperforming a weak sector despite poor absolute results. The clawback connected with a prior accounting correction is evidence that the recovery mechanism was used. Compensation should be evaluated against absolute per-share value, leverage, and through-cycle returns as well as relative TSR.

Celso Goncalves Jr., the CEO’s son, became President and CFO and joined the board in July 2026. His salary increased to $1 million and his change-in-control continuation multiple rose from two to three years; he receives no additional director compensation. The appointment provides continuity and reflects substantial company experience. It also heightens the importance of independent succession oversight. [S7]

A June 5 Form 4 reports that Celso Goncalves sold 214,308 shares in an open-market transaction at a weighted-average $13.4136, leaving approximately 184,542 directly owned shares. The form did not identify the transaction as pursuant to a Rule 10b5-1 plan. Grants, vesting, withholding, and routine transactions should not be mislabeled as open-market purchases or sales; the reviewed current evidence did not identify a comparable open-market purchase at the 2025–2026 lows. [S8]

Management behavior indicates a preference for scale, asset control, and strategic flexibility; the repurchase-then-issuance sequence weakens confidence in per-share price discipline. This is an inference from actions, not a claim about personal intent. [S1][S4][S7]

Verdict: Capital allocation is mixed-to-poor. The 2020 acquisitions were strategically transformative, current debt priority is appropriate, and the 2025 offering protected liquidity. The adverse evidence is pro-cyclical repurchasing, subsequent dilution, an unproven Stelco return, high compensation during losses, and concentrated family leadership. Sustained debt repayment without renewed large-scale M&A is the necessary rehabilitation test.

Changes and Headwinds — Last Two Years

Results have been driven more by external steel prices, trade rules, automotive schedules, and input costs than by internal actions, but contract exit, facility idlings, mix redirection, and cost programs materially amplified the 2026 rebound. [S1][S2][S5]

The business environment changed materially through higher Section 232 protection, evolving derivative-product rules, persistent global overcapacity, improved realized pricing, and a 2026 labor-renewal cycle. [S9][S10][S13][S14]

The core steel tariff increased to 50% in June 2025. Subsequent proclamations refined treatment for derivatives, metal content, and trading partners, making the current regime more complicated than a uniform 50% tariff. Domestic pricing strengthened, but global excess capacity and modern North American investment continued. Policy support and competitive capital formation therefore moved in opposite directions.

Internally, Cliffs exited the loss-making slab contract, idled high-cost facilities and mining capacity, reduced configured raw-steel capacity, sold or marketed non-core property, and redirected production toward hot-rolled, coated, and automotive products. The Q2 collapse in slab volume and favorable product-mix bridge are the clearest filing-level evidence that the operating system changed as management described. [S1][S2]

Stelco moved from acquisition integration toward potential contribution. Management said Canadian hot-rolled operations improved, while downstream finishing remained weaker and required better Canadian trade conditions. Because no current stand-alone financials are disclosed, the magnitude of improvement remains a management claim to test against consolidated margins and future disclosure. [S5]

Middletown changed from a proposed hydrogen-oriented decarbonization project into a defined blast-furnace modernization. The current scope includes the furnace, materials handling, process controls, and cogeneration, with a $500 million DOE award supporting a $1 billion investment. The revised project may offer better near-term economics than an unsupported green premium, but it maintains exposure to integrated steelmaking and final implementation terms remain relevant. [S6][S18]

Leadership changed when Celso Goncalves became President/CFO and a director while Lourenco Goncalves remained Chairman and CEO. Management described the move as an early stage of succession but said the CEO planned to remain for several years. [S5][S7]

Labor is the most immediate operating uncertainty. Management characterized the opening of negotiations as constructive in July. The Steelworkers’ September 4 update said negotiations continued under an extension, acknowledged progress on some issues, stated that significant work remained, and described the company’s response as disappointing. These accounts can coexist, but the later counterparty evidence contradicts any assumption that an agreement was effectively complete. [S5][S9]

No material 2026 accounting-policy change was identified; comparability is affected instead by earlier accrual corrections, Stelco purchase accounting, LIFO movements, impairments, and recurring non-GAAP exclusions. [S1][S2][S4]

Important changes in markets, facilities, and management include protected domestic pricing, slab-contract expiration, upstream idlings, Stelco integration, Middletown modernization, unresolved USW bargaining, and the promotion of Celso Goncalves. [S1][S6][S7][S9]

Verdict: The two-year change set is positive for near-term EBITDA and mixed for long-term value. Pricing, contract exit, and footprint rationalization materially improved earnings. Higher debt, unresolved labor economics, continuing capital needs, and commitment to the integrated route prevent an operating recovery from becoming a demonstrated business-quality upgrade.

Risk Analysis

Factors that could cause a material stock decline include weaker realized pricing, failure to deliver Q3 guidance, poor EBITDA-to-cash conversion, renewed ABL borrowing, labor disruption, lower automotive output, import pressure, major outages, project overruns, or another dilutive financing. [S2][S3][S9][S10]

Leverage creates convex downside. A $300–500 million EBITDA miss matters much more to equity than to enterprise value because interest and maintenance spending do not decline proportionately. Working capital can initially release cash in a downturn, but continued losses eventually consume liquidity. Asset values may support creditor recovery without protecting common shareholders.

A catastrophic investment loss would require a prolonged steel and automotive downturn to collide with high fixed costs, shrinking cash generation, weaker borrowing-base support, unsuccessful asset monetization, and the 2028–2029 refinancing window. [S1][S2]

Under that sequence, the company could face heavy dilution, secured financing, distressed exchanges, asset sales at unfavorable prices, or restructuring. Liquidity makes the chain unlikely in the immediate term, and accelerating EBITDA is a powerful offset. It is nonetheless a credible multi-year tail because enterprise value can survive while the residual equity is severely impaired.

A literal total loss is not the central expectation, but a near-total common-equity loss is plausible through restructuring or extreme dilution after several years of sub-maintenance cash generation. [S1][S2]

The principal upside risk to a cautious position is equally nonlinear. If Q3 and Q4 establish a $2.4 billion-plus annualized EBITDA rate, 2027 contracts add material cash margin, and management directs proceeds to debt, equity value can grow faster than EBITDA because each dollar of debt retired transfers enterprise value to shareholders.

Verdict: Risk is asymmetric in both directions. Near-term liquidity and improving EBITDA reduce immediate distress probability, but debt turns ordinary commodity volatility into an equity-tail risk. The disconfirming evidence to a severe bear case is the current earnings trajectory and maturity runway; the unresolved issue is whether that runway produces permanent debt reduction before the next downturn.

Valuation Discussion

Capital structure and current multiples

At the September 11 close of $12.015 and 570.5 million shares outstanding, equity value was approximately $6.86 billion. June carrying debt was $7.703 billion and cash was $70 million. An attributable enterprise value—equity plus net debt—is therefore approximately $14.49 billion. Standard consolidated enterprise value would add the $214 million noncontrolling interest, but pairing that numerator with Cliffs’ adjusted EBITDA after the company removes NCI earnings would mix ownership bases. [S2][S16]

Reported Cliffs equity was $5.604 billion. Subtracting $1.784 billion of goodwill and $1.063 billion of intangibles produces tangible equity of approximately $2.76 billion. The stock therefore traded near 1.22x reported book and 2.49x tangible book. Tangible book removes acquisition intangibles but remains a poor liquidation value because furnace age, reserve value, permits, environmental obligations, and earning power are not captured consistently.

Trailing adjusted EBITDA is trough-distorted but demonstrates how much recovery is already capitalized. FY2025 adjusted EBITDA of $37 million, less negative $85 million in first-half 2025, plus $381 million in first-half 2026, produces approximately $503 million of trailing adjusted EBITDA. Attributable EV is roughly 28.8x that amount. This is not a sensible normalized multiple; it shows that investors have already abandoned trailing earnings as the valuation denominator.

An illustrative 2026 estimate combines $381 million of reported first-half adjusted EBITDA, management’s approximately $575 million Q3 guide, and an analyst estimate of $625 million for Q4 because management said Q4 should outperform Q3 under the forward curve prevailing in July. The result is $1.58 billion and an EV/2026E adjusted-EBITDA multiple of approximately 9.2x. Q4’s $625 million is not guidance and could be wrong if pricing, volume, or costs change. [S2][S5]

Reverse valuation

The reverse approach is more useful. Dividing $14.49 billion of attributable EV by a 6–7x normalized multiple implies approximately $2.07–2.41 billion of sustainable adjusted EBITDA. That is above the illustrative 2026 estimate and requires a meaningful contribution from 2027 contract renewals, continued utilization, slab exit, and lower costs.

A 6–7x range is an analytical assumption, not a universal steel multiple. Cliffs arguably deserves the low end because of leverage, trough losses, capital intensity, and weaker through-cycle ROIC. It could deserve more if the post-idling break-even proves structurally lower and electrical-steel growth becomes material. It could deserve less if tariff-supported margins prove temporary or refinancing risk rises.

Peer context

Nucor and Steel Dynamics merit higher-quality multiples because they remained profitable through the trough, have stronger balance sheets, and can fund growth and shareholder returns simultaneously. Nucor’s Q2 results included favorable items, and Steel Dynamics is absorbing startup costs in aluminum, so neither peer is risk-free or perfectly comparable. Their ability to remain profitable and liquid through the same cycle is nonetheless relevant evidence that Cliffs’ earnings gap is not purely an industry effect. [S11][S12]

A simple trailing comparison makes Cliffs appear prohibitively expensive because its denominator is depressed. A normalized comparison must instead estimate how much EBITDA survives a cycle and how much capital is required to sustain it. Cliffs’ higher financial leverage gives its equity more upside torque, but it should not be confused with higher business quality.

Bear, base, and bull operating scenarios

The following scenarios are estimates, not company forecasts. They explicitly include revenue, margin, reinvestment, leverage, dilution, and terminal economics.

The bear case assumes current trade protection cannot offset weak demand, domestic competition, or lower contract pricing. EBITDA around $900 million would leave little room after interest and capital spending and could place the equity back on a refinancing/dilution path.

The base case gives credit for slab exit, footprint savings, stable tariff support, and part—but not all—of the claimed 2027 renewal benefit. It assumes no major outage or strike and no large acquisition. The bull case requires strong realized pricing, successful automotive renewals, stable volume, reliable furnaces, disciplined capital allocation, and significant cash conversion.

What the market gets right

The market correctly rejects 2025 as a normalized earnings base. Q2 results and Q3 guidance demonstrate that EBITDA can improve by hundreds of millions within quarters. The market also correctly values liquidity and the absence of an immediate bond maturity. Treating the equity as near-term insolvency would ignore those facts. [S2][S3]

The current price also reasonably anticipates contractual lag. Stronger spot pricing can enter automotive and other agreements over time, and contract exit changes mix. The market is therefore not irrational to price earnings above the 2026 run rate. The question is whether it has capitalized too much of an uncontracted 2027 bridge before cash reaches debt.

Fragile assumptions

The fragile bull assumption is that management’s approximately $500 million contract-reset estimate is incremental to a durable Q3/Q4 base rather than partly overlapping with current price realization. The second is that EBITDA converts to cash after approximately $600 million of interest and $700–900 million of capital expenditure. The third is that labor, outages, and Middletown do not absorb the incremental margin.

The fragile bear assumption is that 2025 represents structural earning power. It does not: the slab contract expired, high-cost assets were idled, realized pricing increased, and Q3 guidance is substantially stronger. A valuation based solely on trough free cash flow would ignore real option value and the company’s liquidity runway.

Own-history valuation claims also require refreshing after large price moves. The share price increased 27.8% from the July baseline while tangible equity declined modestly, moving estimated price-to-tangible-book to approximately 2.5x. Historical percentiles calculated at $9.40 are stale and should not be retained without recomputation. [S2][S16]

Verdict: Current valuation discounts a substantial but achievable normalization. Strong operating momentum is the best evidence against a low valuation, while unchanged year-to-date leverage is the best evidence against capitalizing the full recovery. The decisive denominator is not next quarter’s EBITDA alone; it is sustainable, cash-converting EBITDA after the next contract and steel-price cycle.

Variant Perception

The decisive investor questions are whether Q3 guidance converts to cash, whether 2027 renewals deliver a material portion of the claimed $500 million uplift, whether asset proceeds reduce reported debt, whether labor is resolved without disruptive economics, and whether the post-idling break-even is structurally lower. [S2][S5][S9]

Other questions include Stelco’s stand-alone profitability, the amount of North American gross capacity that becomes net capacity after closures, whether Nippon improves U.S. Steel’s automotive cost position, and whether management avoids acquisitions and repurchases until leverage is clearly repaired.

The apparent consensus is that CLF is a tariff-supported earnings-recovery vehicle constrained by leverage. Q3 guidance makes the near-term inflection difficult to dispute. The differentiated question is therefore not whether earnings improve, but how much of the improvement belongs to common equity after interest, capital expenditure, working capital, labor, and debt repayment.

The strongest bull case combines several levers: slab-contract exit, higher realized pricing, 2027 fixed-contract resets, automotive mix, Stelco normalization, footprint savings, property proceeds, and electrical-steel growth. At $2.8–3.2 billion of adjusted EBITDA, annual free cash flow could exceed $1 billion and debt reduction could create a powerful transfer of enterprise value to shareholders.

The strongest bear case concedes strong Q3 and possibly Q4 results but argues that management is extrapolating a tariff-supported price curve and temporary competitor disruption into durable contract economics just as modern capacity expands. On that view, the company enters the next downcycle with most of its $7.7 billion debt still outstanding and a larger furnace-investment program.

Five load-bearing assumptions decide the outcome:

1. Realized pricing must support at least a high-single-digit to low-double-digit EBITDA margin.
2. EBITDA must convert to cash after interest, maintenance, pension/OPEB, and working capital.
3. Fixed-price renewals must improve realized margin without sacrificing material volume.
4. Labor and furnace reliability must preserve shipments and cost guidance.
5. Trade protection and capacity closures must support domestic spreads despite gross additions.

The factor model is a dated statistical risk diagnostic, not a fundamental classification or causal explanation. As of September 10, it estimated market exposure of 1.48, small-size exposure of 1.18, Materials-sector return exposure of 1.09, and credit-risk exposure of 0.29. Residual volatility was approximately 51.5% annualized, residual momentum was close to zero, residual Sharpe was 0.13, and R² was 31.5%. Most historical return variance therefore remained unexplained by the included factors. The negative daily alpha and high residual volatility support conservative position sizing; they do not prove operational causality. [S17]

Prior transferable assumptions require discipline. The leveraged-cyclical mechanism is confirmed: one positive-FCF quarter improved confidence in the earnings turn before it established cumulative balance-sheet repair. Government support at Middletown lowers capital cost but does not underwrite revenue or utilization. The price gap since July invalidated stale valuation percentiles. Unrelated research mechanisms from biotechnology, banking, marketplaces, restaurants, and utilities do not transfer to this steel investment case.

Verdict: The variant is not that the earnings recovery is fictitious; filing evidence shows it is real. The variant is that enterprise value has recognized the recovery faster than the company has retired debt. The next two quarters determine whether that gap closes through cash generation or through a valuation reset.

Fact vs. Interpretation

Verdict: The highest-confidence facts establish an operating turn and a still-leveraged balance sheet. The most consequential forecasts—contract-reset economics, sustainable margins, asset-sale proceeds, and optional strategic projects—remain management claims or analyst estimates rather than independently demonstrated cash returns. [S2][S5]

Open Questions

1. How much of Q3’s approximately $575 million expected adjusted EBITDA converts to operating cash after receivables, inventory, interest, pension/OPEB, and capital expenditure? [S2][S3]
2. Do expected property sales close, and does net debt fall by the proceeds after taxes and spending?
3. What tonnage, price, surcharge, and volume assumptions support the approximately $500 million 2027 contract-reset estimate? [S5]
4. Does Q4 outperform Q3 without another increase in the steel-price curve?
5. What wage, benefit, staffing, and operating-flexibility terms emerge from USW bargaining? [S9]
6. What are Stelco’s stand-alone EBITDA, maintenance capital, and working-capital requirements?
7. How much announced North American capacity becomes net supply after closures and delays? [S10]
8. Does Nippon’s investment materially improve U.S. Steel’s cost and auto-grade competitiveness?
9. What final milestones, eligible costs, clawbacks, and reimbursement timing apply to Middletown? [S6][S18]
10. How many production programs and annual tons support aluminum substitution?
11. Does POSCO produce a definitive, non-dilutive agreement?
12. What portion of the EBITDA recovery comes from price versus a permanently lower cost base?
13. Will management preserve debt priority if the equity strengthens?
14. What normalized cash tax rate applies after loss carryforwards and jurisdictional effects?

What Must Be True

Bull test

The bullish thesis requires Q3 adjusted EBITDA of at least approximately $550 million, Q4 EBITDA no lower than Q3 if benchmark prices are stable, second-half free cash flow of at least $500 million before optional asset sales, and gross debt moving below approximately $7.2 billion by year-end or soon thereafter. Shipments should remain within the 16.5–17.0 million-ton guide, no material labor disruption should occur, and 2027 realized contract pricing should validate a substantial portion of management’s proposed uplift. [S2][S3][S5][S9]

Bull falsifiers are measurable: Q3 adjusted EBITDA below $500 million; Q4 deterioration despite stable pricing; ABL borrowings remaining near $900 million after announced property receipts; renewed equity issuance for ordinary liquidity; or 2027 realized pricing that fails to improve cash margin. A policy change matters through applicable rates, import volumes, realized prices, and utilization—not through unsupported political prediction. [S2][S13][S14]

Monitoring signals are quarterly realized price per ton, shipment volume, cash margin, operating cash flow, receivables, capital expenditure, ABL borrowing, net debt, fixed-contract disclosures, USW terms, and unplanned outages.

Bear test

The bearish thesis requires the present recovery to prove temporary: realized steel pricing declines, new domestic or imported supply compresses spreads, contract renewals disappoint, labor or outages reduce shipments, and free cash flow fails to retire debt. Under that sequence, Cliffs approaches the June 2028 ABL maturity and 2029 notes without a sufficiently improved leverage profile. [S2][S9][S10]

Bear falsifiers are four consecutive quarters of positive free cash flow, cumulative net-debt reduction above $1 billion, adjusted EBITDA sustained above a $2.4 billion annualized rate across changing spot prices, stable automotive volume after contract resets, and evidence that post-idling ROIC exceeds the cost of capital after maintenance and Middletown spending. [S2][S6][S16]

The extreme view that Cliffs cannot recover EBITDA has already been falsified. The durable concern—that a high-fixed-cost, acquisition-built steel system may fail to convert cyclical EBITDA into acceptable per-share returns—has not. That is the thesis now under test.

Linked primary sources: 2025 Form 10-K, Q2 2026 Form 10-Q, Q2 earnings release, 2026 proxy statement, Middletown announcement, USW bargaining update, and OECD Steel Outlook 2026.

Public source appendix

• S1: Cleveland-Cliffs 2025 Form 10-K — primary SEC filing; published 2026-02-09; Items 1, 1A, 2, 7 and 8; financial statements and Notes 3, 8, 9, 16, 18 and 19
• S2: Cleveland-Cliffs Q2 2026 Form 10-Q — primary SEC filing; published 2026-07-23; Financial statements; debt, liquidity, cash flow, Steelmaking results, product shipments and market-risk tables
• S3: Cleveland-Cliffs Q2 2026 earnings release — primary company disclosure filed with SEC; published 2026-07-23; Financial highlights, adjusted-EBITDA reconciliation, Steelmaking performance, liquidity and Q3 outlook
• S4: Cleveland-Cliffs 2026 proxy statement — primary SEC filing; published 2026-04-02; Compensation Discussion and Analysis; 2025 annual-incentive scorecard; LTI, ownership and clawback disclosures
• S5: Company Financials — Cleveland-Cliffs Q2 2026 earnings-call transcript — management transcript via Company Financials, reconciled to primary filings; published 2026-07-23; July 23, 2026 prepared remarks and Q&A; pricing, Q3/Q4 outlook, 2027 contracts, working capital, property sales, labor and Stelco
• S6: Cleveland-Cliffs announces $1 billion Middletown Works investment — primary company disclosure; published 2026-08-21; Project scope, funding framework, four-year deployment, capacity and completion timing
• S7: Cleveland-Cliffs Form 8-K — promotion of Celso Goncalves — primary SEC filing; published 2026-07-24; Item 5.02; roles, board appointment, compensation and related-party relationship
• S8: Cleveland-Cliffs Form 4 — Celso Goncalves Jr. — primary SEC insider filing; published 2026-06-05; Table I and footnote 1; 214,308-share open-market sale on June 5, 2026
• S9: United Steelworkers Cleveland-Cliffs bargaining update 7 — primary labor-counterparty statement; published 2026-09-04; September 4 status update on extension, progress and unresolved issues
• S10: OECD Steel Outlook 2026 — authoritative industry report; published 2026-06-04; Executive summary, financial conditions and global capacity outlook through 2028
• S11: Nucor Q2 2026 results — primary peer disclosure; published 2026-07-27; Q2 highlights, EBITDA items, segment results, liquidity, capital returns and Q3 outlook
• S12: Steel Dynamics Q2 2026 earnings release — primary peer disclosure filed with SEC; published 2026-07-20; Q2 results, steel operations, cash flow, aluminum startup, capital allocation and three-year after-tax ROIC
• S13: Proclamation adjusting steel and aluminum imports — official policy document; published 2025-06-03; Clauses establishing the 50% additional steel duty effective June 4, 2025 and specified exceptions
• S14: Proclamation further adjusting metal-product tariff regimes — official policy document; published 2026-06-01; Core-metal, derivative, temporary and trading-partner rate distinctions
• S15: Company Financials — Cleveland-Cliffs Q1 2026 earnings-call transcript — management transcript via Company Financials, reconciled to primary filings; published 2026-04-20; April 20, 2026 prepared remarks and Q&A; pricing lag, volume leverage, projects, labor and strategic commentary
• S16: Company Financials — profile, statements, ratios, valuation and daily prices — third-party Company Financials dataset reconciled to primary filings; published 2026-09-11; NYSE:CLF resolved before retrieval; 2021–2026 financial series and prices through September 11, 2026, reconciled to filings
• S17: The factor model — CLF snapshot — internal quantitative diagnostic; published 2026-09-10; Factor exposures, residual signals and diagnostics dated September 10, 2026
• S18: Energy Department announces $500 million award to revitalize American steelmaking — official government project disclosure; published 2026-08-21; Award amount, revised project purpose and $1 billion total Middletown investment