So far this year, US technology companies have announced 123,653 layoffs, representing a 66% increase compared with the same period in 2025. Across all industries, US employers announced 97,006 job cuts in May, up from 83,387 in April. The technology sector recorded the highest number of layoffs among all industries, followed by transportation (6,909 layoffs) and services (6,268 layoffs). Behind these figures lies a puzzling contradiction: many of the same companies carrying out large-scale layoffs are simultaneously committing the largest capital expenditures in their history to AI.

Where Has the Money Gone?

Google, Amazon, Microsoft, and Meta have collectively planned US$725 billion in capital expenditures this year, up 77% from last year. Approximately three-quarters of that spending is directed toward AI infrastructure, including servers, GPUs, and data centers. Meta CEO Mark Zuckerberg explicitly told employees that the approximately 8,000 positions eliminated by the company were a direct result of increased AI infrastructure spending.

Amazon’s capital expenditure plan stands at approximately US$200 billion, Alphabet’s at around US$180–190 billion, and Meta’s at approximately US$125–145 billion. Microsoft has yet to provide full-year guidance, but its AI-related spending alone has already exceeded US$100 billion.

Almost all of this money is being directed toward computing power, custom chips, GPUs, and data centers rather than employee compensation, retail expansion, or share buybacks. Companies are using labor-cost savings to acquire computing assets capable of replacing portions of human labor.

The Industries Facing the Greatest Impact

Official employment data already provides evidence of AI’s growing impact on the information sector.

According to the Bureau of Labor Statistics’ April employment report, the information industry lost 13,000 jobs during the month, including telecommunications (-3,000 jobs), motion picture and sound recording industries (-6,000 jobs), and computing infrastructure, data processing, and web hosting services (-4,000 jobs).

Employment in the information sector has now declined by 342,000 positions, or 11%, since its recent peak in November 2022.

This figure suggests that workforce contraction in the technology sector is no longer reflected only in layoff announcements but is increasingly visible in actual employment data.

ADP’s May employment report, released yesterday (June 3), offered further confirmation: information services lost 9,000 jobs during the month. ADP specifically noted that this “may be the result of AI-driven growth.”

What Should Investors Watch in Today’s Non-Farm Payrolls Report?

Today is Non-Farm Payrolls day, and interpreting the report may prove more challenging than usual.

The US Bureau of Labor Statistics will release the May employment report at 8:30 PM Beijing time tonight.

Last month’s report surprised markets. April Non-Farm Payrolls increased by 115,000 jobs, nearly double the widely expected range of 62,000–65,000, while the unemployment rate remained unchanged at 4.3%.

Wall Street’s consensus forecast for May calls for payroll growth of approximately 80,000 jobs, with the unemployment rate remaining at 4.3%.

Federal Reserve officials will be closely monitoring the report ahead of the June 16–17 monetary policy meeting.

However, there is an important interpretation trap here.

There is often a time lag between layoff announcements and actual employment losses. Announced layoffs do not necessarily result in immediate job separations, and many displaced workers may quickly find new employment.

Notably, unemployment insurance claims have not risen alongside the increase in layoff announcements. This helps explain why broader labor-market data continues to “look strong.”

Market analysts are focusing not only on the headline payroll figure, but also on three specific indicators: whether the unemployment rate rises above 4.3% (a move above that threshold would significantly shift the market narrative), the labor force participation rate (continued declines would undermine the optimistic interpretation of the broader employment data), and average hourly earnings growth (overly strong wage growth could delay expectations for interest rate cuts).

These indicators may ultimately prove more important than the headline payroll number itself in determining how markets respond.