Artificial intelligence is rapidly reshaping corporate America, from cloud infrastructure to semiconductors, yet its economic footprint is not fully visible in official growth statistics. According to a new analysis from Goldman Sachs, the mismatch between corporate revenue and government data leaves a $115 billion blind spot in US GDP reporting.
The AI Boom vs. Government Data
Since 2022, US companies supplying AI infrastructure — from chipmakers to cloud providers — have reported a $400 billion surge in revenue. On the surface, that suggests AI is a major driver of recent economic expansion.
However, Goldman’s analysts estimate that AI has lifted real US economic activity by $160 billion, equivalent to 0.7% of GDP. Only $45 billion (0.2% of GDP) of that has been officially captured in government statistics, leaving most of AI’s contribution missing from headline numbers.
This disconnect stems from the way the Bureau of Economic Analysis (BEA) calculates GDP, particularly its treatment of semiconductors and computing hardware.
Why the Numbers Don’t Add Up
The BEA counts high-performance semiconductors — essential for AI training and deployment — as intermediate inputs rather than final investment goods. In practice, this means:
-
When advanced chips are imported, their cost is subtracted from GDP.
-
Their role in powering AI models is not logged as direct investment.
-
The intangible assets created by AI development — from enterprise solutions to generative models — are largely invisible in GDP data.
Goldman estimates that $75 billion spent on AI model development and enterprise cloud solutions has not been classified as investment in official statistics. This gap underscores how traditional economic measurement struggles to capture the value of intangible, software-driven innovation.
Tariffs and Import Timing Add Distortions
The picture is further complicated by trade policy. In early 2025, business investment in information-processing equipment spiked as companies rushed to import servers and networking gear before President Donald Trump’s tariffs took effect.
Goldman cautions that this “frontloading” effect inflated the appearance of AI-related investment in the short term, while the offsetting impact of imports being subtracted from GDP means the boost never fully shows up in growth numbers.
Corporate America Feels It, But Earnings Don’t Show It
Beyond government statistics, companies themselves are still struggling to quantify AI’s bottom-line benefits. A record share of S&P 500 firms mentioned AI on their Q2 2025 earnings calls, but only a small fraction provided measurable data on its financial impact.
The result: investors and policymakers are left with a paradox. AI is clearly fueling billions in infrastructure spending and reshaping business strategy, yet its measurable impact on productivity and earnings remains modest, at least for now.
The Bigger Economic Question
Goldman’s analysis highlights a broader debate: how should modern economies measure the value of digital infrastructure and intangible assets like AI models? Traditional GDP metrics, designed for a manufacturing economy, risk undercounting the technologies driving the 21st-century economy.
If AI spending continues to grow at current rates, the official blind spot could widen — leaving policymakers with an incomplete picture of economic momentum just as they face decisions on tariffs, interest rates, and labor market support.
