U.S. Updates Section 232 Tariffs on Steel, Aluminum and Copper

The Analysis

Strictly speaking, this is not a new tariff program, and treating it as one is how an importer ends up a week late and a good deal of money out. What came out of the White House on 1 June 2026 is a proclamation that modifies regimes already in force. It modifies the tariffs imposed in Proclamation 9704, as amended, Proclamation 9705, as amended, and Proclamation 10962, as amended, and it modifies the product lists carried in Annex I-A, Annex I-B, Annex II and Annex III of Proclamation 11021. The authority it recites is section 232 of the Act, 19 U.S.C. 1862(d), alongside sections 604 and 301 of title 3 of the United States Code, and the recital leans on the national security findings made in 9704, 9705 and 10962.

The date that does work in this instrument is not 1 June. It is 12:01 a.m. eastern daylight time on 8 June 2026, and it bites on goods entered for consumption, or withdrawn from warehouse for consumption, on or after that moment. Seven days from signature to effect. Short. And the event that triggers the duty is the entry, not the sailing date, not the arrival, not the date printed on the commercial invoice.

I have read enough entry summaries to know the two mistakes this invites. One is reading 1 June as the day the money changed hands and discovering on 8 June that it did. The other is worse, because it stays invisible until somebody does a warehouse withdrawal: cargo that landed on 5 June and went into a bonded warehouse is not grandfathered. Withdraw it on 20 June and you withdraw it into the new regime. I have watched a team congratulate itself for beating a deadline by three days and then pull the goods out of the warehouse three weeks later at the higher rate. This proclamation names warehouse withdrawals in the same breath as entries, twice. Do not read around it.

The summary that circulated with this story used four verbs: it expands 15 percent eligibility, revises U.S. content rules, adjusts scope and adjusts administration. Let me put each of those against the actual clauses, because the mapping is where the useful reading is.

The 15 percent sits in clause (2)(b), and it is not what it sounds like. For products of Argentina, Ecuador, El Salvador, Guatemala, Japan, the Republic of Korea, Liechtenstein, Switzerland, Taiwan, the United Kingdom, or a member nation of the European Union, the rate is determined by reference to the product's current ad valorem, or ad valorem equivalent, rate of duty under Column 1 of the Harmonized Tariff Schedule of the United States. Where that Column 1 rate is less than 15 percent, the sum of the Column 1 rate and the additional section 232 ad valorem rate shall be 15 percent.

Where the Column 1 rate is at least 15 percent, the additional section 232 rate shall be zero percent. Read that again. It is a top-up, not a grant. The figure being quoted around, 15 percent, is a combined result that only exists for goods whose ordinary duty rate sits below that line, and the benefit for goods already at or above it is that the additional duty disappears while the Column 1 duty stays exactly where it was. Zero additional is not zero duty, and I have seen that distinction derail a pricing meeting.

Next to it sits clause (2)(c), which runs the same arithmetic against a different target: 10 percent, determined on the Column 1 rate in the manner outlined in (2)(b), for derivative articles the aluminum content of which is composed entirely of aluminum smelted and cast in the United States, or the steel content of which is composed entirely of steel melted and poured in the United States. Note the two pairs of verbs. Smelted and cast for aluminum. Melted and poured for steel. They are not interchangeable, and the split is deliberate: the two industries describe their first solidification step differently, and the drafters used each industry's own term.

Now the default, and this is the number that will do the damage wherever nothing else applies. Under clause (2), with effect from 8 June 2026 until 11:59 p.m. eastern standard time on 31 December 2027, the additional ad valorem rate for all aluminum and steel articles listed in Annex I-C to this proclamation shall be 25 percent, unless a lower rate applies under clause (2)(b), (2)(c) or (2)(d). That is a temporary regime with its own expiry printed into the text.

Here is the exception I would circle before anyone opens a spreadsheet. Clause (3)(b) provides that if a product is subject to multiple rates of duty under clause (2), the lowest applicable rate shall apply. So these clauses are not a ladder you climb until one fits and then stop. If your derivative article comes from a listed jurisdiction and its steel content was also entirely melted and poured in the United States, you are arguably inside both (2)(b) and (2)(c), and you take the lower construction. Ten beats fifteen. That one sentence is worth more than most of the commentary written about this proclamation, and it is easy to miss, because it sits in the clause about 2028 rather than in the clause about today.

And here is a caution I want stated plainly, because this is the kind of thing that gets repeated until it hardens into a fact. The 25 percent sentence in clause (2) is written for aluminum and steel articles listed in Annex I-C. Copper is inside this instrument: Proclamation 10962 is one of the regimes being modified, clause (3)(a) carries 10962 forward into the 2028 rates, and clause (4) speaks of copper smelted and cast in the United States. But the 25 percent in clause (2) is not, on the face of the text, the copper rate. Do not carry it across. Whatever applies to your copper lines sits in 10962 as amended, and that is a document your broker should pull rather than a number you inherit from a summary.

The U.S. content rules are the other half of what changed, and they are where I would slow down rather than speed up. Clause (4) is a deeming rule. A product's metal content shall be deemed composed entirely of aluminum smelted and cast in the United States, of steel melted and poured in the United States, or of copper smelted and cast in the United States, if such aluminum, steel and copper account for at least 85 percent of the weight of the aluminum, steel, and copper of the product. Strictly speaking, deemed is the operative word. Deeming is a legal fiction: the content does not have to be entirely American, it has to clear a weight threshold, and once it does you treat it as though it were entirely American.

That is generous. It is also a trap, because the denominator is not the product. It is the weight of the metal in the product. A machine that is 60 percent steel by weight, of which 90 percent was melted in the United States, clears the test on the metal pool and fails any intuitive test you might apply to the finished item. Run it on the metal, not on the box.

There is an ambiguity in that sentence I am not going to resolve for you. The phrase is at least 85 percent of weight of the aluminum, steel, and copper of the product. Read one way, the three metals form a single pool and you test the U.S. share against combined metal weight. Read another way, you test each metal separately. The two readings diverge badly for a product with a little copper and a lot of steel where the copper is imported and the steel is domestic. I find the single-pool reading the more natural of the two, but I am not going to build a filing position on my own reading when the Secretary has been directed to have Customs and Border Protection issue guidance on exactly this area. Wait for it.

Canada and Mexico work differently again, under clause (2)(d). For products of Canada and Mexico that qualify for preferential tariff treatment under the United States-Mexico-Canada Agreement, a duty of 25 percent applies only to the non-U.S. content of the product, and non-U.S. content is defined as the total value of the product minus the value attributable to parts produced in the United States. Two things follow. This test is measured in value, not weight, so the denominator has changed from clause (4). And there is a floor: the total effective duty assessed under this subclause shall not be less than 15 percent ad valorem, as detailed in Annex IV.

Then come the two sentences that I think matter more than the rate, and that nobody quotes. The Secretary shall issue guidance to Customs and Border Protection regarding the assessment of U.S. content for the purposes of this clause. Guidance is coming; it is not here. And if CBP determines that an importer has engaged in fraud or has deliberately misled the United States Government regarding the United States content of an imported product, CBP shall impose penalties to the extent permitted by law. So the U.S. content figure is not a box you tick and hope for the best. It is a representation made at entry, with a penalty attached, on a methodology whose details are still being written, and it will be revisited later against records you will have to produce.

Timing has two cliffs, not one, and only the first is visible from here. The regime I have just described runs to 11:59 p.m. eastern standard time on 31 December 2027. With effect from 12:01 a.m. eastern standard time on 1 January 2028, the applicable rates for products listed in Annex I-C are the rates set out in clause (3) of Proclamation 11021. I have not read clause (3) of Proclamation 11021 and I am not going to paraphrase it from memory. What I will say is structural: a contract, a price list or a quotation that runs past the end of 2027 is written against a rate table nobody can price today, and it should carry a rate-change provision rather than a fixed duty assumption. That is a drafting point, not a forecast.

Scope moves too, by a mechanism people forget to watch. Clause (5) directs the Secretary, in consultation with the Trade Representative, the Chair of the International Trade Commission and the Secretary of Homeland Security, to determine whether any modifications to the HTSUS are necessary to effectuate or implement the proclamation and to make them through notice in the Federal Register, and it allows technical corrections to any annex. So the tariff codes that put you in or out of scope are amendable by notice. Your scope screen is a recurring control with an owner and a review date, not a project you finish in July and file away.

Clause (6) tells you to expect more of this. The Secretary is to continue to monitor imports of metal products. The Secretary and the Trade Representative are to review the status of those imports with respect to national security. Both are to inform the President of any circumstance that might indicate a need for further action and, in the other direction, of any circumstance that might indicate that any of these actions are no longer necessary. Clause (10) supersedes inconsistent provisions of earlier proclamations and executive orders. Clause (11) is a severability clause: if any provision, or its application to any individual or circumstance, is held invalid, the remainder stands. I do not read tea leaves on litigation. I will say what a severability clause is for: it is drafted by people who have thought about the possibility that part of the instrument falls.

On the inclusion side, the preamble does two things worth noting. Agricultural equipment and certain HVAC systems and components that are predominantly for residential use shall be included in the category of derivative products subject to the temporarily-reduced 15 percent ad valorem duty under Proclamation 11021. And the tariffs imposed on mobile industrial equipment and machinery are temporarily modified. Both are inclusions into a reduced rate, so the direction is relief rather than burden. Both are described as temporary or temporarily-reduced, so do not build a five-year capital plan on either. Note the shape of the HVAC language as well: predominantly for residential use. That is a use-based qualifier attached to a product category, and use-based qualifiers are where classification disputes live.

Now the arithmetic, with every assumption on the table, because a rate table only becomes real when you run it on your own numbers. Assume an importer of a derivative steel article with entered value of USD 500,000 per month, and assume the Column 1 duty rate on that article is zero. Zero is my assumption, made to keep the arithmetic clean; substitute your actual Column 1 rate, because under clauses (2)(b) and (2)(c) it changes the answer.

Assume the article comes from an origin not on the clause (2)(b) list and the steel was not melted in the United States. Clause (2) default: 25 percent. That is USD 125,000 a month, USD 1,500,000 a year. Assume the same article from Japan, which is on the list, with the same zero Column 1 rate. Under clause (2)(b) the sum of Column 1 and the additional rate is 15 percent, so the additional rate is 15 percent: USD 75,000 a month, USD 900,000 a year. The difference against the default is USD 600,000 a year on these assumptions. Now assume the same Japanese article where the steel content was entirely melted and poured in the United States, so clause (2)(c) is in play and clause (3)(b) lets you take the lower outcome. The combined target becomes 10 percent: USD 50,000 a month, USD 600,000 a year. On these assumptions, melt and pour location is worth USD 900,000 a year against the default case.

I am not telling anyone to move a mill. Resourcing is a commercial decision and it is not mine to make. What I am saying is that at this size, origin and melt location have stopped being footnotes on the entry summary and have become inputs to a landed cost model, and if procurement is still buying against one delivered price with no duty line broken out, the model is wrong.

Run Canada and Mexico separately, because the floor does something non-obvious. Duty on total value is 25 percent of the non-U.S. content share. Write the U.S. content share as s and the effective rate is 0.25 times one minus s, subject to a floor of 15 percent. Solve for where the floor bites: 0.25 times one minus s falls below 0.15 once s exceeds 0.40. So on a USMCA-qualifying product of Canada or Mexico, every percentage point of U.S. content above 40 percent of value buys you nothing. You are pinned at 15 percent. Below 40 percent, each point of U.S. content is worth a quarter of a point of duty. If you are negotiating with a supplier over how much value to place in the United States, that 40 percent line is the number to put in front of them, and it falls out of the proclamation's own two figures rather than out of anyone's estimate.

One more, on the cost of being wrong. Assume one entry line at USD 500,000 where a clause (2)(c) claim is made and the melt-and-pour documentation later fails. The gap between the 10 percent combined outcome and the 25 percent default is 15 points, which is USD 75,000 on that entry. Assume the same claim was made monthly for eighteen months before anyone looked at it. Eighteen times USD 75,000 is USD 1,350,000, before penalties, and clause (2)(d) says penalties run to the extent permitted by law where the government finds fraud or deliberate misleading. That figure is arithmetic on my assumptions, not a legal conclusion about how CBP would assess a given case. I am not going to predict that. I am pointing out that the downside of an unsubstantiated content claim is asymmetric: it costs very little to make and a great deal to have made.

This column is dated 30 June 2026, so the effective date is already behind us and the question is remediation rather than preparation. By 10 July, have the trade compliance lead produce a line-level screen of every entry in the last ninety days against Annex I-C and against the derivative lists as modified, with each line marked in scope, out of scope or awaiting determination, and nothing left blank, reported by entered value. By 24 July, for every line marked in scope, have the broker record the origin, the mill, and whether any U.S. content claim is being made and on what basis. Where the answer is that no claim is being made, write that down too, because silence later reads as an omission rather than as a decision.

Before 30 September, obtain smelt-and-cast or melt-and-pour documentation for the suppliers covering the top 20 percent of in-scope entered value and hold it with the entry records for the full retention period your broker and counsel apply. Assume forty SKUs across twelve mills and three working days per mill to chase and validate, which is thirty-six analyst days; at an assumed fully loaded USD 400 a day, that is roughly USD 14,400 of one-off effort against the USD 900,000 annual differential in the example above. That ratio is the whole argument for doing the paperwork.

The traps, in the order I expect people to hit them. Reading the 15 percent as a rate that was granted rather than a combined outcome that is calculated. Assuming cargo that arrived before 8 June is safe while it is still sitting in a bonded warehouse. Treating the scope screen as finished when clause (5) lets the HTSUS be amended by notice. Carrying the clause (2) 25 percent across to copper when clause (2) is written for aluminum and steel articles listed in Annex I-C. Ticking the U.S. content box because the 85 percent deeming rule makes it easy, and then being unable to produce records when clause (2)(d) is enforced against you. And on stacking: the summary notes this sits alongside existing IEEPA and Section 301 tariff programs. Whether and how these additional duties stack with those on a given HTS line is a question for your broker and your counsel on your actual codes. I am not going to assert a stacking rule I have not verified.

Here is where I would concede the argument to someone who reads it differently. If the CBP guidance defines parts produced in the United States narrowly, the 40 percent crossover moves, U.S. content claims get smaller, and everything I said about the value of documenting U.S. content shrinks with it. If the guidance adopts the per-metal reading of the 85 percent test rather than the single-pool reading, some of the products I described as clearing the test will not clear it. If the rates in clause (3) of Proclamation 11021 turn out to be lower than the temporary regime, then 31 December 2027 is a relief rather than a cliff and the contracts I told you to revisit will not need revisiting. I have not read that clause.

What does not change under any of those readings is the documentation. Every outcome in this proclamation that is better than 25 percent is a claim you have to substantiate, and the substantiation is the part that will still be sitting in your files in 2028. I would rather spend July on mill records than on forecasts.

↑ Back to the key points

— By Dr. Ingrid Voss