Q2 US GDP growth came in at a disappointing 1.5% last week. Still,
AI investment seems to have helped drive up economic growth. Intellectual property products plus information processing equipment, two subcomponents that are closest to the AI investment surge, accounted for 0.67 percentage points of GDP growth.
As the narrative goes, however, such AI-related investment relies on imports—chips and RAM, for example—which are mostly not produced in the US and they have to be imported from countries like Taiwan and South Korea.
Since imports are not part of domestic production, these investments are actually much less growth-enhancing, as they are subtracted by the corresponding import offset.
Is AI’s import-reliant characteristic the reason US GDP was rather weak?
How strong was the offset?
Through the lens of international trade data, we can see that computers, peripherals and semiconductors have been accounting for both an increasing dollar amount and percentage share of total imports.
As we have said, the information processing equipment & IP products’ contribution to Q2 growth was 0.67 percentage points.
To calculate the offset, we use the combined GDP contribution of four import lines—computers, peripherals & parts; semiconductors; telecommunications equipment; and charges for the use of IP—as a proxy for the AI import offset. The subtracting effect was relatively weak in Q2, having a negative contribution of merely 0.2 percentage points.
But if we look at the bigger picture, we can see that the import offset is actually more broad-based outside the AI-related investment.
Equipment investment—a broader category that comprises information processing equipment, industrial equipment, transportation equipment and some other equipment—contributed a substantial 0.8 percentage points, almost completely offset by the negative contribution (-0.73 ppts) of capital goods except automotive imports, which include aircraft; computers, peripherals & parts; semiconductors; telecom equipment; industrial equipment; and electrical equipment.
Moreover, in Q2 we also saw a near-complete offset between consumption goods (PCE goods in the graph below) and consumer goods imports. This is the first time consumer goods imports have had an offsetting effect since Q1 2025, right before Trump’s “Liberation Day” tariffs.
In total, all the import offset added up to over 1.5 percentage points.
Private domestic final purchases (PDFP), which measures only consumption and investment growth to reflect the domestic demand coming from the private sector, grew 3.94%. Import offsets, along with the 0.67 ppt negative contribution from change in inventory, are the two major factors behind Q2’s lackluster growth.
