AI infrastructure frenzy sweeps the US! Surge in computing equipment imports pushes trade deficit to highest level since March 2025
In July, the goods trade deficit unexpectedly widened to $118.8 billion, significantly above the market expectation of $100.5 billion, marking the largest gap since March 2025.
According to Futu News APP, as long-term U.S. Treasury yields continue to rise, global investors focus on the Jackson Hole Central Bank Annual Symposium and U.S. economic data, the U.S. goods trade deficit unexpectedly widened to $118.8 billion in July, far above the market expectation of $100.5 billion, marking the largest deficit since March 2025—a so-called “boom-type trade deficit.” Behind this is the dominance of capital goods imports, which drove total imports up 3.7%, while goods exports fell 2.9%, highlighting the further divergence between U.S. domestic investment demand and foreign trade performance.
From the perspective of asset pricing, if the expansion of the U.S. trade deficit continues, it will drag down real Gross Domestic Product (GDP) in the third quarter via net exports. However, imports of capital goods—especially those related to artificial intelligence infrastructure—continue to show strong growth, indicating that corporate capital expenditure, particularly investments and construction processes closely tied to AI computing infrastructure, remain in full swing.
With Nvidia once again reporting much stronger-than-expected results and an explosive outlook, the hot trading topic around AI computing power is likely to extend far beyond Nvidia GPU clusters. It could accelerate and spread to the entire AI computing power industry chain, touching upon HBM/DRAM/NAND, CoWoS/3D advanced packaging, data center CPUs, optical interconnects, and data center power infrastructure, igniting a new industry-wide “main uptrend” super rally.
Morgan Stanley forecasts that nearly $3 trillion in AI-related infrastructure investment will flow through the global economy by 2028, with more than 80% of expenditures still ahead. According to the latest estimates from Goldman Sachs, global AI capital expenditure is projected to increase from $765 billion annually in 2026 to $1.6 trillion annually in 2031. Cumulative capital expenditures from 2026 to 2031 are expected to reach about $7.6 trillion, and U.S. data center power demand is projected to rise from 31GW in 2025 to 66GW by 2027. This will directly push AI computing infrastructure investment into server CPUs, DRAM/NAND/HBM, advanced packaging, liquid cooling, power equipment, transformers, gas turbines, grid connection equipment, data center REITs, and engineering construction and other segments.
Driven by the rebound in the transportation of capital goods into the U.S., imports surged, pushing the July goods trade deficit to its largest level since early last year.
Data released Thursday by the U.S. Department of Commerce shows the goods trade deficit widened by 17.2% over the previous month to $118.8 billion, the largest since March 2025. Prior institutional economist surveys had a median forecast of only a $100.5 billion deficit. These data are not adjusted for inflation.

U.S. goods imports in July grew more than expected by 3.7%, while exports of goods decreased 2.9%. Imports of capital goods (mainly computers and accessories, semiconductors, and telecommunications equipment) saw the largest increase since 1993.
Over the past few months, the trade deficit has fluctuated, due to factors such as the war in Iran driving up global demand for U.S. crude oil and refined petroleum products. At the same time, U.S. companies have been stockpiling goods and raw materials under energy inflation pressures to guard against any negative impacts from supply chain disruptions. As imports of AI-related infrastructure hardware remain robust, companies are also adapting to shifting tariff policies.
The $118.8 Billion Deficit Highlights AI Boom! The U.S. Is Forging an AI Computing Power Empire Through a Surge of Imports
In July, U.S. goods imports rose 3.7% month-over-month to $318.2 billion, with a strong rebound in capital equipment transport, while goods exports fell 2.9% to $199.4 billion. This is highly consistent with the underlying trend of U.S. AI data centers continually importing high-performance AI server components, AI computing accelerators, network infrastructure devices, data center power systems, and liquid cooling components. However, the Commerce Department’s preliminary data have not yet released detailed AI-related import numbers, and changes in tariffs, supply chain stockpiling due to the Iran war, and other capital goods imports also amplified the deficit.
Cross-referenced with Nvidia’s financial reports, the evidence for the booming U.S. AI computing infrastructure buildout is even more complete. In fiscal Q2 2027, Nvidia’s revenue surged 106% year-over-year to $96.2 billion, with data center revenue up 117% to $89 billion. Third quarter revenue guidance stands at $108 billion, with a 2% range of variation; management further expects FY2028 revenue to grow around 70%, emphasizing that this outlook is still constrained by supply capacity, and the Vera Rubin platform has entered full mass production ramp-up.
In other words, Nvidia’s financials validate the “order and compute demand” side, while capital equipment imports corroborate that “chips, servers, and data center infrastructure are being transported cross-border and converted into physical assets within the United States.” Further research from the Minneapolis Fed shows that by 2025, AI-related products will account for 23% of U.S. imports, up 73% from 2023, whereas non-AI imports only grew 3% over the same period. Without the AI infrastructure boom, the U.S. goods trade deficit in 2025 could be nearly $200 billion less.
Retail inventory data released alongside the goods trade figures showed inventories increased by 0.7%. Wholesalers continued to build up stocks.
Trade and inventory data will assist the government in making a preliminary estimate of third quarter GDP, which will be released in October. Prior to the latest goods trade report, the Atlanta Fed’s GDPNow model projected that net exports would lower GDP growth by 0.14 percentage points. By comparison, net exports dragged down GDP growth by 1.14 percentage points in Q2.
The more comprehensive July trade data, including the service trade account balance and inflation-adjusted goods trade figures, will be released on September 3.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
Hormel Foods Trims Full-Year Sales Guidance Amid Weak Consumer Demand
Dollar General, Dollar Tree Raise Earnings Outlook as Tariff Refunds Boost Quarterly Bottom-Lines
Nvidia (NVDA.US) stages a "Back to the Future" comeback: 70% growth guidance shocks Wall Street, analysts rush to upgrade, bullish sentiment in options market surges
Nvidia's latest quarterly results once again far exceeded market expectations, proving its unshakable dominance in the AI chip sector. This directly drove its stock price to surge in pre-market trading and prompted several analysts to raise their target prices.

Jackson Hole has not yet begun, but the Federal Reserve’s “hawkish” signal comes first! Schmid bluntly says interest rates are still not tight enough, raising rates risk returns to the trading desk
Kansas City Federal Reserve Bank President Jeff Schmid stated that as inflation continues to remain above the 2% target, the central bank's interest rate policy has not curbed the US economy.

