A 7.5% overcapacity tariff on Chinese goods: what has been reported, and what is actually in force

What was actually reported

On 24 August 2026, Bloomberg reported that the administration is preparing an additional 7.5% tariff on imports from China, to be issued under the Section 301 investigation into structural excess capacity and production in manufacturing sectors that USTR opened on 11 March 2026. Officials were said to be aiming to publish the investigation’s results before President Trump and President Xi meet in Washington on 24 September 2026.

The sourcing is “people familiar with the matter”. Bloomberg’s own sources cautioned that exact rates have not been finalised, and described one option in which higher headline rates are announced with portions suspended so the effective rate lands at 7.5%. That is a meaningful detail: a suspended tariff is a lever that can be pulled later without a new investigation.

So the status, as at 25 August 2026, is precise and worth stating plainly:

• No USTR determination has been published in the excess-capacity investigation.
• No Federal Register notice, no product list, no HS codes, no effective date.
• No US official has confirmed either the 7.5% figure or the 20% ceiling it is calculated from.

Why 7.5%, and not some other number

The interesting thing about 7.5% is that it was not derived from any finding about excess capacity. It is arithmetic.

On 27 July 2026, China’s Ministry of Commerce disclosed for the first time that Washington had committed, during bilateral consultations, to cap its replacement tariffs on Chinese goods at 20%. Beijing traced the commitment to the Kuala Lumpur consultations of October 2025 that produced the one-year truce. In the same statement it noted that the current replacement rate is 12.5% — which is another way of saying that 7.5 percentage points remain unused, and that Beijing has counted them.

A tariff sized to fill a ceiling to the millimetre is a negotiating position wearing the clothes of a trade remedy. CSIS made the point sharply on 10 August 2026: a credible economic answer to Chinese excess capacity “could easily far exceed 7.5 percent”, so a 7.5% outcome is evidence that the truce, not the investigation, is setting the number. The same analysis flags the legal exposure — Section 301 was written to address specific practices of specific partners, and uniform low-rate tariffs applied across almost everything invite the argument that the statute is being used for something else.

Note also that the ceiling is Beijing’s account of a private commitment. The US has not confirmed it. If Washington disputes the framing, 7.5% stops being the obvious answer.

Where 20% sits inside what you actually pay

This is the single most common misreading, and it is worth being blunt about. “Twenty per cent on China” is not the duty on your goods. The 20% ceiling covers one layer of a stack. Underneath it, everything else is still running.

Add those together and the picture is not 20%. The trade-weighted average duty on Chinese-origin goods has been running near 30%, and specific lines run far above it. Brookings found that only about 16% of 2025 US imports from China were free of both the 2018-era Section 301 tariffs and the newer duties — so for roughly five goods in six, the reported 7.5% would be an increment on something, not a starting point.

Two consequences follow. First, an extra 7.5 points on a good already carrying 25% plus metals duties is a smaller proportional shock than the headline suggests. Second, there is no way to know your number from an article. It comes from your HS code. See how the tariff stack works, and have a licensed customs broker confirm before you re-price anything.

Three tariff regimes in six months

The reason a 7.5% tariff is being routed through an overcapacity investigation, rather than simply announced, is that the previous route was closed by the Supreme Court.

Read as a sequence, this is a story about authority rather than about China. Each replacement has been narrower, slower and more procedurally demanding than the last — an investigation, a comment docket, a hearing, a published determination. That is worse for anyone who wants tariffs quickly and better for anyone who needs to plan, because Section 301 leaves a paper trail with dates on it.

What “overcapacity” means, and why the word is contested

Overcapacity means an industry that can produce more than its market can absorb, with the surplus pushed into export markets at prices domestic producers cannot match. USTR’s position is that structural excess capacity in several economies is unreasonable and burdens US commerce. The evidential anchor is not subtle: China posted a record goods trade surplus of nearly US$1.2 trillion in 2025, up from US$992 billion in 2024 — achieved even though its exports to the United States fell 28% over the same year.

Beijing’s answer arrived on 28 July 2026 as a formal position paper from the Ministry of Commerce, released in Chinese and English. Its argument is that a surplus is not a synonym for excess capacity:

• Large, durable surpluses are normal for major manufacturing economies. Germany’s and Japan’s current-account surpluses have exceeded 6% of GDP; China’s runs at about 3.7%.
• Foreign-invested enterprises produced 27% of China’s exports in 2025 and 16% of the surplus — so a meaningful share of what is described as Chinese overcapacity is foreign firms manufacturing in China.
• Capacity utilisation is set by global demand and the international division of labour, not by domestic subsidy alone. No single government, on this reading, has standing to declare another’s capacity excessive.

Both accounts can be partly right, and for a buyer the argument is largely beside the point. What matters is that the definitional dispute is unresolved, which is exactly why the tariff can be sized by negotiation rather than by economics — and why the same question keeps resurfacing in Brussels, Delhi and Brasília whenever Chinese exports are redirected away from the US market. Buyers outside the United States should read the whole episode as a signal about where the next trade-defence case lands, not as somebody else’s problem.

Which products are in scope

Nothing is in scope yet, because no list exists. What USTR published on 11 March 2026 was an illustrative list of sectors it considers affected by structural excess capacity:

Aluminium · automobiles · batteries · cement · chemicals · electronics · energy goods · glass · machine tools · machinery · non-ferrous metals · paper · plastics · processed food and beverages · robotics · satellites · semiconductors · ships · solar modules · steel · transportation equipment.

That is twenty-one sectors covering most of what a general importer buys, which is why the sector list is close to useless for planning and the HS code is everything. Note too that the forced-labour tariffs already carve out several categories — goods already subject to Section 232, USMCA-qualifying imports, certain raw materials and manufacturing inputs, and some agricultural goods. If the overcapacity action follows the same drafting pattern, exclusions will do a great deal of the real work, and they will be written at code level.

What changes for a buyer, and what does not

Arithmetic first. A 7.5% ad valorem duty adds US$75 to a US$1,000 customs value — and it is charged on the customs value, not on your retail price, so the effect on margin is smaller than the percentage sounds and the effect on cash flow at entry is immediate. The importer of record pays it. Not the supplier, not the marketplace, whatever the quotation says.

What genuinely changes:

• Cash at the border, before revenue. Duty is payable at entry. For a small importer running on a thin working-capital line, an extra 7.5 points on a container is a financing problem some months before it is a margin problem.
• Quotes with a longer shelf life than the rules. A 90-day price validity written in July can be crossed by a tariff announced in September. If your terms do not say who carries a change in duty between quotation and entry, they should.
• Incoterms that were fine become expensive. Under DDP the seller carries duty; under FOB or EXW you do. Both are defensible. Not knowing which one you signed is not. See our Incoterms guide.
• Smaller buyers absorb more. This is the least discussed and most reliable effect. Large importers re-engineer, split origins and negotiate; a business importing four containers a year does none of that and pays the difference.

What does not change: your supplier’s cost base, your product’s classification, your quality problems, or whether the company you are buying from actually manufactures anything. A tariff moves a number. It does not move any of the things that decide whether a China supply chain works.

Money already owed to importers

Here is the part of this story with a deadline attached, and it has nothing to do with the 7.5%.

After the Supreme Court struck down the IEEPA tariffs, the Court of International Trade ordered CBP to refund what had been collected — on the order of US$165 billion. CBP built a consolidated mechanism, CAPE, inside the ACE portal, and opened it in phases:

By early August 2026 the administration said it had refunded around US$100 billion, roughly 60% of what was collected. Two things are worth knowing. Refunds are not automatic — the mechanism relies on the importer coming forward with entry summaries and proof of payment. And on 3 June 2026 the Department of Justice appealed the CIT’s refund orders to the Federal Circuit, arguing relief should not extend to importers who never sued. Importers who did not file protective actions face delay, and on some categories the risk of permanent loss.

Moving production, and why it just got harder

The instinctive response to any China tariff is to look at Vietnam, Malaysia, Mexico or India. Sometimes that is right. On a 7.5-point move it usually is not: qualifying a new factory, re-tooling, re-certifying and absorbing the first year of quality variance routinely costs more than the duty, and the comparison people run in a spreadsheet almost never includes the second year of defects.

More importantly, the arbitrage has narrowed. In August 2026 the White House published a report naming more than 40 countries as routes through which Chinese-origin goods are disguised, put the revenue loss at US$19–26 billion a year, and described an AI-assisted screening programme to flag suspect routing at the border. The US–Vietnam framework already carries a punitive rate for goods judged to be transshipped, at double the ordinary rate.

The rule underneath all of it has not changed and is worth stating exactly: origin follows substantial transformation, not the shipping address. A container that stops in a third country, or is repacked there, or has a label applied there, is still Chinese-origin. Real relocation — where the transformation genuinely happens elsewhere and the records show it — remains entirely lawful and is a legitimate strategy. Paperwork describing a transformation that did not happen is fraud, and it is now being looked for by machine.

What Beijing can do, and the date that matters more

China has said it rejects the overcapacity characterisation, calls the investigation unilateral, and reserves the right to respond. No specific countermeasure has been announced against this plan, which is consistent with a tariff calibrated to stay inside a ceiling Beijing itself disclosed. The tools used in previous rounds are known: tariffs on US exports, additions to the unreliable entity list, export controls on critical materials, procurement pressure, and WTO complaints that take years.

But the date to have in your calendar is not 24 September. It is 10 November 2026.

That is when the truce agreed in October 2025 expires. On that day China’s suspension of its expanded rare-earth and critical-minerals export controls also lapses, and the suspended tranche of retaliatory tariffs comes back into view. A 7.5% tariff inside the ceiling is, on this reading, a modest event deliberately staged before a much larger one. If you are modelling risk, model 10 November — and note that a September summit is where both sides find out whether there is a deal to extend it.

There is a quieter counter-current worth knowing about, because almost nobody has noticed it. Following the May 2026 Trump–Xi meeting the two governments established a US–China Board of Trade, and USTR ran a docket on up to US$30 billion of Chinese-origin goods for tariff relief, on a “30 for 30” basis with reciprocal Chinese reductions. Comments closed on 10 July 2026 and rebuttals on 27 July 2026. Tariffs are being added and subtracted at the same time, by the same agency, in the same quarter. Anyone describing this period as a one-way escalation is not reading the dockets.

Quick FAQ

Has the US imposed a 7.5% overcapacity tariff on Chinese goods?

No. As at 25 August 2026 no such tariff exists. Bloomberg reported on 24 August, citing people familiar with the matter, that one is being prepared under the Section 301 excess-capacity investigation opened on 11 March 2026. Its sources said exact rates were not final. There is no determination, no Federal Register notice and no effective date.

Why is the figure 7.5%?

Because it is the headroom. On 27 July 2026 China’s Ministry of Commerce said Washington had committed to cap replacement tariffs on Chinese goods at 20% and noted the current rate is 12.5%. The gap is 7.5 points. The number comes from the ceiling, not from a finding about excess capacity — which is the most informative thing about it.

Would that make total US tariffs on Chinese goods 20%?

No. The ceiling covers one layer: the replacement tariffs imposed since 2025. Beneath it sit the MFN rate, the 2018 Section 301 tariffs at 7.5% or 25%, Section 232 metals duties, and any anti-dumping or countervailing duties. The trade-weighted average has been running near 30% and individual lines run much higher. Your number comes from your HS code.

When would it take effect?

Unknown. Officials were reported to be aiming to publish the investigation’s results before the 24 September 2026 Trump–Xi meeting in Washington. Publication of findings is not the same as an effective date, and Section 301 actions normally arrive with a Federal Register notice that sets one.

Which of my products would be covered?

No product list has been published. USTR’s illustrative sector list runs to twenty-one industries, from steel and solar modules to processed food and robotics, which is broad enough to be unhelpful. Scope will be decided by HS code, and exclusions — as in the forced-labour action — will do much of the real work.

Should I move sourcing out of China?

Rarely on a 7.5-point move alone; qualifying a new factory usually costs more than a year of the duty. Concentration risk is the better argument for a second country. If you do move, note that origin scrutiny tightened sharply in 2026 and origin turns on substantial transformation, not on where the goods were shipped from.

Can I recover the tariffs the Supreme Court struck down?

Possibly, and it is not automatic. CBP’s CAPE mechanism opened on 20 April 2026 and refunds are being paid, but the government has appealed the scope of the refund order and importers who never filed protective actions at the CIT face delay or permanent loss on some categories. Liquidation status decides it. Take this to a licensed customs broker or trade counsel now rather than at a deadline.

Does any of this affect selling into China?

Not directly — a US import tariff is paid by the US importer and has no bearing on the Chinese market. The indirect route is retaliation, and the date to watch is 10 November 2026, when the current truce and China’s suspension of its expanded critical-minerals export controls both expire.

What would change this article?

A USTR determination in the excess-capacity investigation; a Federal Register notice with product codes and an effective date; a US statement confirming or denying the 20% cap; the outcome of the 24 September meeting; or the expiry or extension of the truce on 10 November 2026. This page carries a last-checked date at the top for that reason.