U.S. Consumer Spending Remains the True Engine: Why AI Investment Impact on GDP is Overstated

AI is Not the Primary Engine of the 2025 U.S. Economy: Consumption Remains the True Pillar

For some time, a prevailing narrative has suggested that Artificial Intelligence (AI) is the central engine sustaining the U.S. economy. However, recent research indicates that this narrative is significantly overstated.

Undeniably, the AI boom has profoundly influenced market valuations, catalyzed massive capital inflows, and driven record-breaking bond issuances to finance data centers. Particularly in early 2025, the impact of AI investment on Gross Domestic Product (GDP) was prominent, leading many economists and market participants to conclude that AI spending “saved” an otherwise sluggish economy.

However, Prajakta Bhide, U.S. Economic Strategist at MRB Partners, argued in a January report that the primary driver of U.S. GDP growth in 2025 remained personal consumption—the standard hallmark of economic expansion—while AI-related capital expenditure (CapEx) ranked second.

“AI is a significant part of the growth story, but it isn’t the whole story,” Bhide stated in a recent interview with CNBC. “The narrative that U.S. GDP would have contracted last year without AI CapEx is simply not supported by the data. The U.S. consumer continues to be the primary engine of this expansion.”

Import Leakage Dampens AI’s Actual Contribution to GDP

Bhide points out that because a substantial portion of high-tech equipment is imported, the net contribution of AI investment to GDP is far smaller than intuition suggests.

GDP is comprised of four components: Personal Consumption, Investment, Government Spending, and Net Exports. Since GDP measures only domestic production, imports are excluded from the final calculation.

Her research found that, without accounting for imports, AI-related investment contributed an average of approximately 90 basis points (0.9 percentage points) to real GDP growth between Q1 and Q3 of 2025, representing about 40% of the average GDP growth during that period.

However, once AI-related imports—including computers, components, semiconductors, and telecommunications equipment—are deducted, the net contribution of AI investment drops to approximately 40–50 basis points, accounting for only 20%–25% of real GDP growth. Furthermore, Bhide noted that despite data center construction dominating headlines, the most significant AI-related GDP contributions in 2025 actually stemmed from software and computer equipment rather than physical infrastructure.

“Even if AI-related optimism were to face a shock, the more realistic, smaller net estimate of AI’s contribution debunks the popular claim that the economy would stall without it,” Bhide wrote in her January 8 report. “Even without the AI boom, GDP growth in 2025 would have remained above 1.5%, largely supported by robust personal consumption.”

Institutional Consensus: The Economic Scale of AI is Exaggerated

Bespoke Investment Group echoed this sentiment in late 2025, noting that specific structural anomalies in the first quarter led the market to severely overestimate AI’s share of the economy. The firm found that in the second and third quarters of 2025, categories associated with AI spending accounted for only about 15% of quarterly GDP growth and represented less than 5% of total GDP.

While final annual data for 2025 has yet to be finalized, quarterly figures present a complex picture. Real GDP grew at a robust annualized rate of 4.3% in Q3, following a 3.3% expansion in Q2. Conversely, Q1 saw an annualized contraction of 0.3%, marking the first quarterly decline since early 2022.

Looking Toward 2026: Consumption as the Critical Support

Bhide’s research reinforces the central role of household consumption in economic expansions. Looking toward 2026, she expects U.S. consumption to remain resilient, even as income growth moderates and wealth continues to concentrate among high-income earners.

She noted that fiscal policy will continue to provide a degree of support to the broader economy, partially offsetting the slowdown in income growth. “Overall, the U.S. consumer is in good shape,” Bhide remarked.

Addressing concerns that consumption is fragile due to a perceived reliance on the wealthy, Bhide remains unconvinced. She argued there is insufficient evidence that consumption is being “hollowed out” or that this represents a significant cyclical risk.

For 2026, she anticipates that U.S. economic growth will be supported by multiple factors, including sustained AI investment, the Federal Reserve’s rate-cutting cycle, and a stabilizing unemployment rate amid a significant decline in immigration. Moving forward, her focus remains on shifts in productivity data and the pace of employment growth.