Strong Jobs Data Triggers Nasdaq’s Worst Day in a Year — And Exposes the AI Capex Paradox

The US economy added 172,000 jobs in May — more than double the 85,000 expected. Unemployment held at 4.3%. And the stock market crashed. The Nasdaq dropped 4%, its worst day since the tariff turmoil of early 2025. Nvidia, AMD, and Micron led the selloff. The 10-year Treasury yield spiked above 4.5%. The 30-year breached 5%.

Good news is bad news. The economy is too strong for rate cuts. And that creates a paradox at the center of the AI buildout.

The AI Capex Paradox

The AI infrastructure buildout runs on capital. Microsoft committed $80 billion in capex. Meta guided $125-145 billion. Alphabet just raised $80 billion in equity. Nvidia’s customers collectively spend $300+ billion annually on GPU clusters, data centers, and networking.

All of this spending is financed — either through debt, equity raises, or retained earnings that could have been returned to shareholders. When interest rates rise, the cost of financing increases. A 30-year yield above 5% means every data center loan, every corporate bond issued for AI infrastructure, and every lease agreement costs more to service.

The paradox: the AI buildout needs a weak economy (low rates, cheap capital) to be funded at the scale the industry demands. But AI adoption accelerates in a strong economy where businesses have revenue to spend on AI tools. Today’s jobs data pushed rates higher, making the capex more expensive, while simultaneously confirming that businesses are hiring and have budget for AI adoption.

Why Chip Stocks Got Hit Hardest

Nvidia, AMD, and Micron dropped more than the broader market. Three reasons:

Valuation sensitivity. Chip stocks trade at 20-30x forward revenue. At those multiples, small changes in the discount rate (driven by Treasury yields) cause large changes in present value. A 50 basis point rise in the 10-year yield can compress Nvidia’s fair value by 10-15% purely through the valuation math.

Broadcom aftershock. Broadcom’s AI revenue miss on Wednesday is still reverberating. The market is questioning whether AI infrastructure revenue can grow fast enough to justify the capex. Today’s rate spike compounds that doubt — now the capex is more expensive AND the revenue growth is uncertain.

Rotation to safety. Investors moved into healthcare (J&J +2%), staples (Colgate +3%, Coca-Cola +3%), and away from growth. This isn’t AI-specific skepticism — it’s the classic risk-off rotation when yields spike. But AI stocks, as the highest-multiple segment, have the farthest to fall.

The SpaceX IPO Timing

SpaceX is in the middle of its $75 billion roadshow. Pricing is June 11 — six days away. Today’s selloff is the worst possible backdrop. Institutional investors who would allocate to SpaceX are watching their tech portfolios decline 4% in a single day. The appetite for a $1.77 trillion tech IPO at 95x revenue diminishes when the 30-year yield is above 5% and the Nasdaq just had its worst day in a year.

This doesn’t kill the SpaceX IPO. But it may force a pricing adjustment. If SpaceX prices below $135/share, the entire AI IPO narrative — Anthropic at $965 billion, OpenAI at $852 billion — gets repriced downward.

The Global Contagion

South Korea’s Kospi dropped 5.5% — its worst day of the year. South Korea’s economy is structurally tied to the semiconductor cycle through Samsung and SK Hynix. When chip stocks sell off in New York, Seoul follows within hours. The AI trade is now large enough that a single US jobs report can move markets across the Pacific.

What This Means for the AI Economy

One day does not change the structural thesis. The $300 billion in committed AI capex is contractual — it doesn’t get unwound because of a single jobs report. Microsoft, Google, and Meta have signed data center leases and GPU purchase agreements that extend through 2028.

But the cost of that commitment just went up. And the market’s willingness to fund new AI ventures at trillion-dollar valuations just went down. The AI economy was built on the assumption of cheap capital and infinite growth. Today, for the first time in this cycle, both assumptions got tested simultaneously.

The AI capex paradox won’t resolve in a day. But the companies that survive it will be the ones generating enough operating cash flow to self-fund their infrastructure — not the ones depending on public markets to write $75 billion checks at 95x revenue. That distinction is about to matter a lot.

For the full structural map of the AI economy, read The Map of AI Redrawn on Business Engineer.