July 2026 U.S. Computer Imports Have No Country Split in the Trade Release
July 2026 U.S. trade data show a $14.8 billion rise in computer-related imports, but not which countries supplied it. Bilateral deficits cannot fill that gap.
Those numbers invite an obvious sourcing question. Which countries shipped the additional equipment?
The monthly trade release does not answer it. Its commodity table reports the technology surge at the world level. Its country table reports total goods trade with each partner. The two tables are not a country-by-commodity matrix.
A reader cannot use the release alone to assign the computer surge to Mexico, Vietnam, Taiwan, China, or any other supplier. Import teams need a separate official data pull before changing a sourcing assessment, tariff model, or country-risk assessment.
The aggregate deficit masks the computer-import shift
On a seasonally adjusted Census basis, capital goods excluding automotive products reached $140.264 billion, up $14.396 billion from June. Computers, computer accessories, and semiconductors together increased by $14.757 billion. Their combined rise exceeded the $11.961 billion increase in Census-basis goods imports because declines elsewhere offset part of the gain.
Yet the total goods and services deficit through July was 29.6 percent lower than a year earlier. The aggregate balance and the composition of imports are sending different signals. The first shows a smaller year-to-date deficit. The second shows a sharp increase in imported computing equipment. Traverse's June trade-data analysis explained why July was only a partial post-July 24 tariff month and could not settle whether a sourcing shift would last. The new problem is that the headline tables do not identify where the product increase came from.
Two official tables answer different questions
Exhibit 8 of FT-900 organizes imports by principal end-use category and commodity. It shows the monthly value and change for computers, computer accessories, semiconductors, telecommunications equipment, and other capital goods. These are national totals. The table has no country column.
Exhibit 19 in the July 2026 FT-900 country tables organizes goods trade by selected country and area. It shows that the seasonally adjusted goods deficit with Mexico reached $27.5 billion in July. The deficits with Vietnam, Taiwan, China, South Korea, and Malaysia were $23.3 billion, $18.1 billion, $15.2 billion, $10.4 billion, and $4.8 billion, respectively.
That table has no computer row.
The record-country highlights in the same date-fixed report show deficits with Mexico, Vietnam, Taiwan, Thailand, South Korea, and Malaysia. Those figures are not seasonally adjusted. The record deficit with Mexico was $26.3 billion in the highlights, while Exhibit 19 reported a seasonally adjusted $27.5 billion. Both figures can be correct. Neither identifies Mexico's contribution to the computer increase.
Why bilateral deficits cannot identify suppliers
It is tempting to infer that the largest bilateral deficits identify the suppliers behind the technology surge. They do not. A bilateral balance includes every goods category and moves with both imports and exports. Mexico's total can change because of vehicles, machinery, electronics, agricultural products, or a decline in U.S. exports. Taiwan and Vietnam may have substantial technology shipments, but their balances alone do not measure their contribution to the July increase.
The cross-tabulation exists outside the release. The U.S. Department of Commerce monthly country-by-product trade datasets support import queries by country and end-use code as well as by country and Harmonized System code. The end-use query can test which partners moved within the same economic categories reported in FT-900. The HTS query can then show the tariff classifications inside that movement. The U.S. International Trade Commission DataWeb partner and product query guide provides another route to official merchandise statistics by partner, month, value, quantity, and product classification.
An FT-900 end-use label is an economic grouping, not a tariff line. A detailed customs query is generally unadjusted, while the headline commodity changes are seasonally adjusted. Country, product, entry date, customs value, and quantity answer different questions. This analysis diagnoses what the release itself permits. It does not claim to have completed the separate country decomposition.
What import teams should do now
Start with the official end-use dataset. Pull July and June general imports by end-use code and country for computers, computer accessories, and semiconductors. Keep the data unadjusted. Rank the partner contributions to the unadjusted country-by-end-use movement, then reconcile them to the unadjusted world total. Do not present that result as a decomposition of the adjusted headline change.
Then move to HTS-level records for the leading partners. Compare July 2026 with July 2025 before using the June movement. Compare quantity only within the same 10-digit HTS provision and quantity unit. A higher value may reflect more units, a different product mix, or a higher unit value.
Company records are the final check. Match the public result against HTS number, origin, manufacturer, entry date, customs value, quantity, Chapter 99 treatment, and calculated duty. Supplier headquarters and port of shipment are not substitutes for origin. A large bilateral deficit is not evidence of transshipment, as Traverse's Detective Border analysis showed.
Do not turn the $14.757 billion increase into a duty estimate. FT-900 explains that general import value excludes U.S. duties, freight, insurance, and other charges incurred in bringing merchandise into the country. Nor does its semiconductor category identify products covered by a particular tariff action. Traverse's background on the semiconductor policy stack shows why classification, specifications, end use, and entry documents still control.
What would change the calculus for country attribution
A country-by-end-use pull could show that one or two partners accounted for most of the unadjusted July movement. That result would justify a targeted supplier, origin, and tariff review. A broad increase across many partners would weaken a single-country explanation, but it would not identify the cause. Simultaneous shipment timing, inventory building, product launches, unit values, and tariff timing would remain possible.
HTS and quantity evidence could change the reading again. Concentration in a narrow set of classifications, accompanied by a jump in units, would look different from higher customs values spread across many products. Company entry records may also diverge from the public pattern because a firm's suppliers and entry timing are not the national average.
What the headline tables do not establish
FT-900 establishes that computer-related capital-goods imports rose sharply in July. It does not establish who imported them, where each product originated, which HTS provisions were involved, or what duty treatment applied. Its headline combines goods and services, while Exhibit 19 covers goods. The release cannot prove that tariffs caused the increase, that one partner displaced another, or that the imports all served the same end use.
July's country allocation remains unresolved in the release. Import teams should not fill that gap with a bilateral-deficit ranking. The work requires official country-by-end-use data, HTS detail, and company entry records.
From reading to review
Run the numbers on your lane.
The duty calculator runs the current stack for any HTS code and origin. A free account opens full tool output, AD/CVD detail, Chapter 98 processing, and available exports.