- AI capex by the five largest hyperscalers is on track to top $800 billion in 2026 and $1.1 trillion in 2027 — equivalent to 3.3% of GDP, exceeding projected U.S. defense spending.
- The AI economy grew an estimated 31% in Q1 while the non-AI economy grew just 0.1% — but much of that AI growth is imported silicon, meaning it contributes far less to domestic output than headline numbers suggest.
- S&P 500 earnings are on track to rise 27% in Q1, but the Magnificent Seven alone account for 61% gains while the other 493 companies average 16% — and labor’s share of business output has fallen to 54.1%, its lowest since records began in 1947.
- Contrary to conventional wisdom, an AI investment bust might not devastate the broader economy — just 33 counties account for 72% of U.S. data centers, limiting geographic spillover, and ordinary workers depend more on wages than on the asset prices AI is inflating.
What Happened?
WSJ chief economics commentator Greg Ip argues that AI has moved beyond being a tailwind for the U.S. economy and is now distorting virtually every major economic indicator. AI spending is inflating GDP while suppressing real wage growth. It’s widening the trade deficit (because AI hardware is largely imported) while boosting Taiwan’s trade surplus to an almost unthinkable 24% of GDP. It’s turbocharged Magnificent Seven profits while leaving the rest of corporate America behind. And it’s driving consumer gloom about jobs even as actual layoff filings remain subdued — because the fear of AI displacement is suppressing workers’ willingness to demand raises.
Why It Matters?
Ip’s analysis cuts through several popular narratives simultaneously. The “AI is saving the economy” story is true — but net of imports and geographic concentration, less true than the headlines imply. The “AI bust would be catastrophic” story is also overstated: data centers are concentrated in 33 counties, stock and profit gains accrue mostly to capital owners, and average workers’ livelihoods depend far more on wage growth. Meanwhile, the political economy is toxic: AI lifts asset prices and executive confidence while depressing worker sentiment and wage bargaining power — exactly the dynamic that fuels populist backlash and policy blowback like the emerging compute tax debate.
What’s Next?
The key question is whether AI spending translates into broad productivity gains that eventually lift wages — or whether it remains a narrow phenomenon enriching capital at the expense of labor. If the latter, the political pressure for compute taxes, antitrust action, and AI regulation will intensify. Watch also for whether the AI capex boom sustains itself: at $800 billion-plus annually, it is now large enough that any slowdown would register clearly in GDP data, creating its own self-reinforcing narrative.
Source: The Wall Street Journal
