Cleveland-Cliffs Inc. CLF

Key data

The business

Cleveland-Cliffs is the largest maker of flat-rolled steel in North America, and it is built end to end: it mines its own iron ore, turns it into steel in blast furnaces and electric arc furnaces, and sells the finished sheet mostly to automakers. That vertical setup is the moat. About 40% to 45% of its flat-rolled steel goes out under fixed-price contracts, so the price is locked in for a year at a time and the volatile spot market only touches the rest. Feedstock comes from its own ground, and its Toledo plant makes a low-carbon metallic called HBI that stretches its hot-metal supply. The whole thing sits behind a Section 232 tariff wall that keeps cheaper foreign steel out. It is a price-taker with a domestic floor under it, and a contract book heavy in one cyclical customer: the car.

The numbers

The story is a loss narrowing quarter by quarter, not a company that earns steadily.

The inflection is Q2 2026: revenue hit the biggest quarter in years while the loss more than halved from a year earlier. Against the full fiscal record, the size of the hole is clear.

This is a business that earned $5 a share in the 2021 steel spike and lost $2.91 two years later. The swing is the point: operating leverage this large means a thin move in the price per ton is the whole result. The engine of the turn is margin, not volume.

Gross margin crossed back above zero in June for the first time in over a year, and the operating loss nearly vanished. What the market does not get for free is the leverage on the way up: consensus already models a full swing to about $0.63 of earnings in 2027, which is why the stock at $12.50 costs 20 times that hoped-for number and more than 60 times the cash it earns today. The edge here is not that the stock is cheap. It plainly is not. The edge is whether margins overshoot that modest recovery if auto demand and the tariff floor both hold, and the single print that settles it is Q3 gross margin staying positive.

Management

Lourenço Goncalves built this company by buying: AK Steel, ArcelorMittal USA, Stelco. The debt is the receipt for those deals, and it is now the whole risk. What he did with his own money says more than the deals do. In February 2026, in the middle of a loss-making stretch, he sold $37.3 million of stock, plan status not disclosed, and his son sold another $2.9 million in June; against that, one insider bought $200,000. Capital allocation carries the same warning: the company spent $733 million buying back stock in 2024 at prices above today's, then bought nothing in 2025 as the losses deepened. That is buying high and stopping at the bottom, the opposite of the discipline a levered cyclical needs.

How it fails or surprises you

The debt breaks before margins normalize. Net debt is $7.6 billion, more than 12 times trailing cash earnings, against $70 million of cash. If steel rolls over before the recovery sticks, coverage tightens fast. The cushion is a $3.0 billion undrawn credit line and a covenant not currently in force; the print that flashes first is operating cash flow, which burned $325 million in Q1 2026.

The recovery was one quarter. June's positive margin rests on a single print, and it has not yet met its next test. Q3, due in October, is the quarter that says whether the turn was real or a spot-price blip. The prior watch on exactly this line is still open, not answered.

Auto and tariffs both hold, and margins overshoot (right tail). Fixed contracts reprice annually; if 2027 auto volumes recover and the tariff floor holds, locked-in prices step up and the operating leverage runs the other way. Consensus at $0.63 for 2027 barely prices this. A single strong contract-season print, visible in realized price by early 2027, would reset the earnings power well above where the Street sits.

Closing thoughts

The market has already bought the recovery. At $12.50 the stock discounts the swing from a $1.5 billion loss to a small 2027 profit, which is why every backward-looking multiple looks absurd and the forward one looks merely full. Someone competent is on the other side, betting the auto cycle and the tariff floor both wobble; the edge is small. Your read beats theirs only if you believe the operating leverage on a levered integrated mill is underpriced, because a business that swings from minus $2.91 to plus $5.36 does not resolve into a tidy $0.63. The left tail is fatter today: $7.6 billion of debt against $70 million of cash means a stalled recovery is a solvency question, while the upside is a re-rate you are partly paying for already.

The bet is still that steel prices in America stay high enough, held up by tariffs on foreign metal, for Cliffs to sell the steel it makes for cars at a profit before the next slump. What breaks it is the debt meeting a downturn before the margin turn is proven. Watch two numbers together: Q3 gross margin, and operating cash flow. If the margin holds above zero and the cash burn stops, the recovery is real; if June was the peak, the $7.6 billion is the only story left.