The Long Downturn: How American Blast Furnace Steelmaking Came Apart

On February 6, 2024, all four members of the U.S. International Trade Commission voted that imported tinplate was not injuring American producers. The Commerce Department had found the opposite six weeks earlier, calculating dumping and subsidy margins against Canada, China, Germany, and South Korea. But an antidumping order requires both findings, and without the injury determination the duties evaporated. Cleveland-Cliffs announced within days that it would indefinitely idle its tinplate mill at Weirton, West Virginia, and issued WARN notices to roughly 900 workers. The plant went quiet that April.

The company called the ruling a travesty. The United Steelworkers said unfair trade was choking out the domestic industry. Both were describing something real, and neither was describing the whole of it. Weirton had opened in 1909 and spent a century producing tinplate; the mill had a coke plant, a four-strand continuous slab caster, and a vacuum degassing facility, and by most measures it was not a technological relic. What killed it was not one vote in Washington. It was the last in a sequence of pressures that had been narrowing the ground under integrated American steelmaking for sixty years.

American steel output peaked in 1973 at roughly 137 million net tons. Employment had already begun sliding from its postwar high of about 650,000 mill workers. Today the industry employs somewhere in the neighborhood of 85,000 people in iron and steel mills, and the number of operating blast furnaces in the United States has fallen from about 140 in the 1970s to twelve. No one has built a new one in decades.

That collapse did not begin with a trade dispute. It began with a technology gap that opened while the American industry was at its most profitable.

Rebuilt abroad, unmodernized at home

Austrian engineers commercialized the basic oxygen furnace in the early 1950s. It made steel in under an hour where an open hearth needed eight or more, at lower cost and with less labor. Continuous casting arrived alongside it, eliminating the ingot-and-reheat step entirely. Japanese and West German producers, rebuilding from rubble with no sunk capital to protect, adopted both technologies as a matter of course. American producers, who owned functioning open hearths that were still earning money, adopted them slowly and piecemeal.

The gap became visible during the 116-day strike of 1959, when domestic customers turned to imported steel and found it acceptable. That year, for the first time, the United States imported more steel than it exported, and it never reversed. Washington responded with the first voluntary restraint agreements in 1968, the trigger price mechanism in 1978, and another round of restraints in 1984, three decades of managed trade that slowed the erosion without addressing what caused it.

The minimills

The more damaging competition turned out to be domestic. In 1969 Nucor opened a small mill at Darlington, South Carolina, that melted scrap in an electric arc furnace. It skipped the entire front end of integrated steelmaking, no ore, no coke ovens, no blast furnace, no sinter plant, and it cost a fraction as much to build.

Minimills began at the bottom of the product range, rolling rebar and light bar that the integrated producers were content to abandon, and worked upward into merchant bar and structural shapes. The decisive move came in 1989, when Nucor commissioned a thin-slab caster at Crawfordsville, Indiana, and gained access to flat-rolled sheet. That was the high-margin core of the integrated business, and the product line a tinplate mill depends on.

The advantages compounded. Minimills went up on greenfield sites, mostly in the nonunion South, carrying no retiree obligations. Their smaller scale let them throttle back in a downturn, while an integrated mill running at partial capacity bleeds money, a blast furnace cannot be run at half output, and banking or relining one costs tens of millions. Electric arc furnaces now account for about 70% of American steel production, and every new plant announced in the United States is an EAF.

Moving on

While minimills took market share, the older producers were paying for their own history. Pension and retiree healthcare obligations accumulated across generations of employment left them carrying per-ton costs their competitors never incurred. Capital that might have funded modernization had gone instead to dividends, diversification into unrelated businesses, and debt service.

Youngstown Sheet and Tube closed its Campbell Works on September 19, 1977, a date the Mahoning Valley still calls Black Monday, and eliminated roughly 5,000 jobs in a single announcement. The pattern accelerated. Between 1998 and 2003 some forty American steel companies entered bankruptcy, including LTV, Bethlehem Steel, and National Steel. The Pension Benefit Guaranty Corporation absorbed their pension plans in what remain among the largest claims in its history, and the assets were consolidated into International Steel Group and later into Mittal and Cleveland-Cliffs. Weirton followed the same path by an unusual route: when National Steel divested the plant in 1982, the workforce bought it outright in 1984, in what was then the largest employee stock ownership plan in the country. The mill turned a profit for the rest of the decade. But employee ownership left no outside parent to absorb losses when conditions turned, and the company filed for bankruptcy in 2003.

Steel also lost markets it had held for a century. Aluminum cans took over beverage packaging in the 1970s and 1980s. Plastic containers displaced steel in food, chemicals, and household goods. Automakers moved toward aluminum and composites for weight reduction, and the steel that remained in a car body was thinner and stronger than what it replaced, meaning fewer tons per vehicle. For a mill built specifically to make tinplate, this was not a cyclical problem. The market it existed to serve was structurally smaller than it had been.

Global overcapacity is the pressure that gets the most attention, and it is real. Chinese steel production went from roughly 15% of the world total in 2000 to more than half. Much of that capacity was built with state support and answers to employment and industrial policy goals rather than to price signals, so it keeps running through downturns when a market-disciplined producer would cut output. Estimates of surplus global capacity run in the hundreds of millions of tons, more than the entire American industry produces.

Dumping is the narrower legal concept: selling an export below its home-market price, or below the cost of production. American trade law addresses it through a two-key system. The Commerce Department determines whether dumping or subsidization occurred and calculates the margin. The ITC separately determines whether the domestic industry is materially injured or threatened. Both must be affirmative for duties to issue.

That structure is where the Weirton petition failed. Cliffs and the USW filed in January 2023. Commerce announced duties on four countries in January 2024. The ITC then found no injury, unanimously, and the case ended. Can makers and food processors had argued against the duties, since tinplate is an input cost for them, and the Commission’s injury analysis weighs the health of the domestic industry against evidence in the record, including whether a producer’s difficulties trace to imports or to other causes. A century of the pressures described above supplies plenty of other causes.

Even successful cases have limits. Antidumping duties are country-specific and product-specific, so exporters shift production to a country not named in the order, or alter a product just enough to fall outside its scope. Enforcement becomes a continuing series of circumvention findings and new petitions. This is why Section 232 has become the preferred instrument: imposed in 2018 at 25% on national security grounds rather than injury grounds, raised to 50% in June 2025, and restructured in 2026 into tiered rates that reach derivative products and are assessed on full customs value. It applies broadly and does not require proving harm case by case. It also raises input costs for every American manufacturer that buys steel, which is the tradeoff that trade policy in this sector has never escaped.

Where the twelve stand

The surviving blast furnaces are concentrated in Indiana, Michigan, Ohio, and Pennsylvania, and they produce roughly a third of American steel. The recent news has cut both directions. Nippon Steel completed its acquisition of U.S. Steel in June 2025 for about $14.9 billion, with a pledge of roughly $11 billion in domestic investment by 2028 and a national security agreement whose protections for specific plants begin expiring in 2027. U.S. Steel restarted the idled B furnace at Granite City, Illinois, in late March 2026 after White House and union pressure, bringing back some 500 workers at a plant it had moved to wind down six months earlier, and announced plans to restart tin production at its Gary Works tin mill. In Canada, Algoma Steel made the final charge of its No. 7 blast furnace in early 2026 and completed its conversion to electric arc steelmaking.

That is the shape of the thing. Blast furnace steelmaking in the United States is not being killed by any single decision. It is being outcompeted from below by scrap-based mills, undercut from abroad by capacity that does not respond to price, squeezed by obligations incurred when the industry was dominant, and eroded by materials that took its markets. Trade enforcement can slow the process, and sometimes has. It has never been able to reverse it, because most of what it would have to reverse never had anything to do with trade.