Divergent Compute.AI Economic Think Tank

Publication · the tell

Record results, record cuts.

Nearly every company here posted record or near-record revenue — and cut thousands of people anyway. The stated reason is almost always "AI." So we line the cuts up against the two numbers that never make the layoff press release: where the stock went, and how much the company is actually spending on AI.

2026 YTD · the shape of the cut

128,028
US tech jobs cut · 2026 YTD (Challenger)
$725B
Big-4 AI capex guidance · +77% YoY
40%
of cuts cited AI by May · up from 7% in January
247
tracked events · 200+ companies

The capex line is the one to hold next to the cuts — record AI spend and record cuts, at the same companies, in the same year. See the capex side in Capex Watch →

Tech jobs cut, by month

May was the worst month since Aug 2024.

Q1 ran 66% above the same period in 2025.

Jan
22.3k
Feb
11k
Mar
18.7k
Apr
33.4k
May
38.2k
Jun
4.4k

Biggest 2026 cuts · click a company for its read

A handful of decisions drive the headline.

Five mega-events account for over 60% of YTD volume. Gold = AI cited as the reason.

Oracle
20kAI cited
Amazon
16kAI cited
Intel
15kAI not cited
Nokia
14kAI not cited
Dell
11kAI cited
Meta
10.2kAI cited
Kyndryl
10kAI cited
IBM
9kAI cited

Stock up. Workforce down.

What's really driving the cuts — AI, or the story of AI?

Some names sit at highs and cut anyway — Cisco roughly +82% on the year, Alphabet ~+115%. Others post record revenue yet trade down — Meta, Microsoft, Intuit — and every one of them is cutting. When the stated reason ("AI") rises from 7% of cuts in January to 40% by May, the reason starts to look like a narrative as much as a cause. That gap — between the story and the math — is what Capex Watch measures.

Sources: Challenger, Gray & Christmas (cuts); company filings/guidance (capex). Figures are 2026 YTD and labelled by whether AI was cited — not assumed. A fuller, live tracker is being built.