$155.18
$283.42
bear $155.18
base $229.50
bull $283.42
margin of safety $125.00
market price* $411.17
Reliance is the largest metals service-centre company in North America. It buys carbon steel, aluminium, stainless and alloy products in bulk from mills, cuts and processes them, and resells them in small, urgent orders to more than 125,000 fabricators and manufacturers. In 2025 about 40% of orders were delivered within 24 hours and over 90% came from repeat customers; its US locations shipped about 17% of the industry's tons. Because it buys and sells on the spot market it can pass metal-price changes through, so its gross margin has stayed in a fairly narrow band, and it has been profitable every year since its 1994 listing, including 2009 and 2020. Its profits still rise and fall with metal prices and industrial demand. How we measure earnings. We start from operating cash flow and strip out the swings in receivables, inventory and payables, which in this business are huge and simply track the price of metal: in 2021 working capital absorbed $910 million as prices soared, then released cash later. We also deduct stock compensation and the capital spending needed to keep the plants running (we treat only a small slice of capex as growth spending, tied to sales growth). On that basis owner earnings were roughly $440 million (2016), $519 million (2017), $686 million (2018), $729 million (2019), $559 million (2020), $1,518 million (2021), $1,831 million (2022), $1,136 million (2023), $777 million (2024) and $738 million (2025). Our previous figures did not remove the working-capital swings, which made 2021 and 2025 look far worse than they really were and 2022 better; they are not comparable with these. Intangible amortisation is small (about $40 million a year) and is not a material wasting-asset issue. The anchor. This is a cyclical business, so we use the full 2016-2025 cycle, which contains a trough (2016, 2020), a once-in-a-generation boom (2021-2022) and a normalisation (2023-2025). Adding back roughly $50-80 million a year of cash interest to measure the business before its lenders, the ten-year median is about $830 million, and the two most recent normal years (2024 and 2025) sit at about $790-815 million. We set the base starting level at about $770 million, slightly below both, because 2024-2025 still carry some residual benefit from higher post-2021 metal prices and we do not want to lift the level on conditions that may reverse. Owner earnings grew about -5% in 2025 versus 2024 on a total basis (roughly flat per share); year one of the base case is set at that level with no growth, and growth only begins afterwards. What we credit as growth. Reliance has acquired small service centres year after year for three decades, funded from its own cash and through downturns, in a market where it still holds only 17% share. That is a proven, repeatable process, so the base case credits about 2% a year for ten years, then 1% a year for the following decade, while charging about 20% of cash flow each year as the reinvestment that growth requires; after twenty years we assume no further growth. We do not credit market-share gains from new plants beyond that pace, higher metal prices, or the outsized buyback pace of 2024, since none of these is secured. Debt. At 31 March 2026 the company owed about $1.69 billion: $500 million of 2.15% notes due August 2030, $250 million of 6.85% notes due November 2036, a $400 million term loan due August 2028 and the balance on its $1.5 billion revolving credit facility, which runs to September 2029. Cash was $250 million, of which we treat about $150 million as surplus. The 2030 notes carry an unusually low coupon, which reduces the true cost of the debt by roughly $55 million in present-value terms. Net debt is therefore about $1.49 billion, or about $29 a share. The company is investment-grade, and no maturity looks difficult to refinance. Base case: business cash flow starts around $770 million, grows about 2% a year for a decade and 1% a year for the next as the acquisition programme continues, less reinvestment. The business is worth about $13.4 billion; after $1.49 billion of net debt the equity is about $11.9 billion, or about $230 a share. Bear case: within the next two years a metals and industrial recession of the 2009 kind hits. In 2009 the industry's tons and prices fell sharply at the same time, and Reliance's earnings fell by well over half; in 2020 the shutdown cut 2020 owner earnings to roughly $560 million. We model business cash flow falling from about $760 million to about $300-320 million for two years, recovering to $480 million and then settling at $600 million, the level of the pre-2021 cycle, with no growth and no acquisitions ever after. Even in the trough the cash flow covers interest of about $75 million four times over, so no dilution or forced refinancing is needed. The business is worth about $9.6 billion; after net debt the equity is about $8.1 billion, or about $155 a share. Bull case: the business starts at about $860 million, between the 2024-2025 run rate and the 2023 level, as the new processing plants built in 2023-2025 (capital spending ran well above depreciation in those years) earn their return, and the acquisition programme runs at about 4% a year for a decade (close to its historical pace) then 2% for a decade, with 35% of cash flow reinvested to fund it. The business is worth about $16.2 billion; after net debt the equity is about $14.7 billion, or about $283 a share. What would make this wrong. The single biggest judgement is the normal earnings level. If the 2021-2022 boom permanently reset metal prices and margins higher, our level is too low; if 2024-2025 still flatter the business and the true normal is nearer the 2017-2019 range of $500-730 million, the base case is too high, which the bear case partly captures. Second, the growth credit depends on Reliance continuing to find and integrate acquisitions at sensible prices; it has done so for thirty years, but a large, overpriced purchase would hurt. Third, tariffs and trade policy on steel and aluminium move domestic prices sharply in both directions and are outside the company's control. Margin-of-safety price. We are reasonably confident of the earnings level because ten years of figures, through two troughs and a boom, cluster around $750-830 million once the metal-price swings in working capital are stripped out, and the company has never had a losing year. The worst plausible event is a long, deep industrial recession combined with falling metal prices, which would halve cash flow for two or three years; the structure of the business survives it comfortably, because its debt is modest, long-dated and mostly fixed, it has no customer or supplier dependence, and its assets are mostly inventory and receivables that turn into cash as activity falls. If we are wrong about both the normal level and the acquisition engine, the shares are worth roughly the bear case, around $155. The price assumes a three-to-four-year holding period for the price to reflect value, during which the company pays a dividend of about $4.80 a share and steadily shrinks its share count. From $125 a share, reaching the base value of about $230 within three to four years, plus dividends, implies roughly a 20% annual return, and even the bear outcome of about $155 implies a modest gain rather than a loss. That combination of a bear value above the margin-of-safety price and a credible base-case path to 20% is why the margin-of-safety price is $125 and not higher.