AI Drove U.S. Q2 Growth-But the Import Leak Is Big Enough to Fog the Signal

U.S. Q2 GDP Slowed to 1.5% as Imports Offset Domestic Demand

The second-quarter GDP reading came in at 1.5%, below the 2.1% forecast and down from 2.1% in the first quarter. The important point is not that the U.S. economy suddenly broke. It is that imports rose, and because imports subtract from GDP, they offset part of the strength in consumer spending and business investment.

Why the revision matters

Bulls can still point to a resilient demand profile: consumer spending recovered, and investment was supported by the AI buildout. Bears, however, have the cleaner near-term argument. When growth slows while imports rise, it often means domestic demand is leaking abroad rather than building sturdier domestic momentum.

That is why the next BEA updates matter. The revised estimate will be released in late August, with a final estimate due in late September. If those updates show imports mattered less than feared, the bullish reading regains support. If they reinforce the advance estimate, investors will have less reason to treat 1.5% as evidence of resilience.

The Import Offset Is Broader Than the AI Hype Cycle Suggests

The market may still be underestimating a simple point: even if AI demand is real, the GDP payoff depends on where that spending ultimately lands once orders become shipped equipment.

Trade data already shows the leak

The US trade deficit widened sharply in May, underscoring how much of this buildout is passing through global supply chains. When AI buyers pull in servers, networking gear, and related equipment, the spending can support foreign manufacturers and logistics networks long before any durable domestic earnings payoff appears. In GDP accounting, those imports subtract from growth.

The scale of AI spending magnifies the effect

This is not a marginal issue. The five largest U.S. cloud and AI infrastructure providers are targeting $660 billion to $690 billion in 2026 capex, nearly double 2025 levels. Morgan Stanley then raised its broader estimate to $1.2 trillion to $1.4 trillion in 2028 AI capex. At that scale, even a moderate import share can create a meaningful offset. Big spending does not automatically translate into big domestic value added.

Cash-flow pressure keeps the debate alive

The bullish case is straightforward: this spending will create AI capacity, then revenue, then margin expansion. The cautious view is that many benefits leak through the supply chain before they show up in U.S. earnings power. Recent cash-flow trends make that tension clearer. Amazon reported negative cash flow, Alphabet's cash flow turned negative, Tesla reported negative cash flow, and Meta's cash generation fell 91%. That does not prove the AI thesis is wrong. It does suggest investors are still funding capacity before the revenue and productivity payoffs are fully visible.

What Would Confirm or Challenge the Cautious View

After a 1.5% Q2 GDP print, the more useful question is whether the spending surge is building domestic earnings power or simply financing more imported equipment.

The cautious view weakens if the late-August revision and September final estimate show imports mattered less than feared, or if spending begins to translate into cleaner domestic cash conversion. Right now, the funding burden is still heavy: Amazon reported negative cash flow, Alphabet reported negative cash flow, and Meta's cash generation fell 91%.

For investors, the key test is not whether AI demand exists. It is whether that demand is creating domestic productive capacity and durable earnings power fast enough to offset the wider import leak.