Trade & Sourcing
Who Grew as China’s 2025 Import Value Fell? The Country Totals Do Not Trace the Orders.
In the matched 57-family 2024–2025 panel, China imports fell 17.31%; Taiwan rose 8.20%, India 5.18% and Japan 1.45%. Mexico was nearly flat and Vietnam fell 12.24%. These observations do not prove orders moved between countries.
Editorial evidence cutoff: September 9, 2026. Published September 29, 2026. Observation periods are stated throughout; older figures are retrospective evidence.
Taiwan, India and Japan gained customs value in the tracked manufacturing import panel during 2025. China and Vietnam declined, while Mexico was almost unchanged. The country totals provide a more uneven picture than a simple story in which imports leaving one source all arrive from another.
This retrospective comparison uses the same 57 product families in 2024 and 2025, as preserved in the annual trade records available by September 9, 2026. For September sourcing decisions, it provides a map of historical changes to investigate. It does not follow individual purchase orders, identify interchangeable goods or establish that buyers moved production from one country to another.
THE COUNTRY TOTALS MOVED IN DIFFERENT DIRECTIONS
Taiwan's customs value rose from $28.82 billion to $31.18 billion, an increase of 8.20%. India's rose from $13.95 billion to $14.68 billion, up 5.18%. Japan's rose from $42.43 billion to $43.04 billion, a smaller 1.45% increase from a larger starting base.
China's value fell from $72.11 billion to $59.62 billion, a decline of 17.31%. Vietnam declined from $19.03 billion to $16.71 billion, down 12.24%. Mexico's $99.58 billion became $99.48 billion, a decline of just 0.10% after rounding. Each country had observations across all 57 families in both annual comparisons. The result is not constructed by taking whichever countries happened to appear in a short list of leading suppliers.
A COUNTRY CAN GAIN SHARE WITHOUT GAINING DOLLARS
The full tracked panel fell from $576.27 billion in 2024 to $559.30 billion in 2025. That changing denominator matters. Mexico's customs value was nearly flat, yet its share of the panel increased from 17.28% to 17.79%. It lost fewer dollars proportionally than the overall panel, so its relative weight increased.
Taiwan gained on both measures. Its dollar value increased and its panel share rose from 5.00% to 5.58%. China's share declined from 12.51% to 10.66% alongside its dollar decline. These are different patterns, even though all can appear in a discussion of changing sourcing shares. An article that reports only shares can make a stable dollar flow look like an expanding market without explaining the denominator.
For a sourcing team, the distinction affects the question being asked. Absolute dollar growth describes a change in recorded value from an origin. Share describes its position within the selected panel. Neither directly describes the available capacity, delivery performance or competitiveness of a potential supplier. Those require evidence at a different level.
THE ANNUAL TOTALS DO NOT TRACK A TRANSFER OF ORDERS
The USITC DataWeb records can show that one country gained while another lost. They cannot establish that the gain consisted of the same goods previously purchased from the declining source. A dollar of additional electronics imports is not automatically a substitute for a dollar of lower metal-product imports simply because both belong to the panel.
Even a shared product-family label would leave room for differences in specifications, quality, customers and unit values. Identifying substitution requires more closely matched products and evidence linking the changes. A company deciding whether a new supplier can replace an existing one must also assess technical qualification, delivery and commercial requirements that the country totals do not contain.
The country measure itself follows trade-statistical definitions. Census defines import country of origin within that reporting system. An origin label does not by itself reveal the ownership of the producing company or the location of every upstream input. Country diversification in the aggregate therefore cannot be assumed to eliminate common ownership, component or production-program exposures.
NOMINAL VALUES CAN CHANGE WITHOUT AN EQUAL VOLUME SHIFT
All of the country changes here are measured in customs dollars. Prices, quantities and composition can move at the same time. A country can gain value because its mix shifts toward more expensive goods, while another can lose value without an equally large decline in physical shipments. The annual totals cannot separate those possibilities on their own.
This is particularly important when interpreting the size of China's decline. The 17.31% result is a decline in the recorded value of this defined manufacturing-family panel. It is not a 17.31% decline in every category of U.S. imports from China, or a directly measured loss of that share of physical production. The population and measure remain part of the finding.
The same restraint applies to Taiwan's 8.20% increase. It establishes historical value growth within the panel, but it does not prove that every Taiwanese industry gained U.S. customers. A national total can combine expansion in some families with contraction in others. The next explanatory layer is the country-by-family table, not an assumption that the national result applies uniformly.
A WINNER LIST IS NOT A SUPPLIER SHORTLIST
A list of countries with positive growth can be useful for deciding where to investigate. It is insufficient for deciding where to buy. An existing customer's requirements may have little connection to the product families responsible for a country's aggregate increase. A source showing a decline may still contain suppliers well suited to the particular product.
The historical Mexico result is a useful check against overstatement. Its customs value was almost unchanged, even as its share rose and other origins moved more sharply. A narrative requiring all alternative sources to be rapidly expanding would miss this large, relatively stable flow. Vietnam's decline supplies another check: the measured pattern was not a uniform increase across every country outside China.
WHAT SEPTEMBER SOURCING RESEARCH CAN ACTUALLY USE
For September 2026 planning, the country comparison offers three separate pieces of evidence: absolute scale, historical value change and relative panel share. A useful assessment keeps all three visible, then tests whether the relevant products and qualified suppliers resemble the aggregate pattern. More recent company evidence can show whether an earlier country trend persisted or changed.
The September 9 archive does not turn the 2025 comparison into current monthly news. Its value is structural and diagnostic. Taiwan, India and Japan recorded gains, China and Vietnam recorded declines, and Mexico remained close to its earlier value. The mixed result challenges a simple relocation narrative. It supplies a more specific research agenda: identify the products, distinguish dollars from quantities and establish an actual supplier relationship before calling a country-level change a transfer of business.
Sources and evidence
Evidence period: 2024 to 2025, 57 matched tracked families. The frozen evidence record lists the source files and verified hashes available September 9, 2026. Source revision: 5e4fb7726c3d40060c0151c086c903baae856cab. Later live-data updates do not alter the historical evidence in this article.
census.gov/foreign-trade/guide/sec2.html
Published 2026-09-29.