Epic Games became the week’s highest-profile casualty, laying off over 1,000 employees on March 24. The Fortnite maker’s second major round of cuts in three years wasn’t blamed on AI—but a new CFO survey shows that, across the economy, AI-attributed job cuts are about to accelerate sharply.
Here’s what’s happening in the AI job market this week.
Epic Games: 1,000 Jobs Gone, No AI Excuse
Epic Games announced the cuts after Fortnite engagement declined significantly in 2025. CEO Tim Sweeney didn’t invoke AI as justification. Instead, he acknowledged the company was “spending significantly more than it’s making.”
The layoffs, combined with $500 million in cost savings from reduced contracting and marketing, aim to stabilize the company. Around 200 positions were cut in Cary, North Carolina, where Epic is headquartered.
Severance includes at least four months of base pay (more based on tenure), six months of healthcare coverage, and accelerated stock vesting through January 2027.
This is Epic’s second major round—they cut 830 jobs (16% of workforce) in September 2023.
CFO Survey: AI Job Cuts Will Be 9x Higher
A National Bureau of Economic Research working paper surveying 750 U.S. CFOs dropped a revealing data point: AI-attributed job cuts are expected to be nine times higher in 2026 than in 2025.
But here’s the catch: these same CFOs admit AI isn’t actually delivering productivity gains yet.
The researchers invoked “Solow’s paradox”—the observation that computers appeared everywhere except in productivity statistics during the 1980s. We may be seeing a replay: companies are cutting jobs based on AI’s potential, not its proven performance.
“It’s not the doomsday job scenario that you might sometimes see in the headlines,” said John Graham, co-author and director of the Duke CFO survey. The study found less than half (44%) of surveyed CFOs plan any AI-related cuts at all.
Goldman Sachs senior economist Ronnie Walker noted there’s still “no meaningful relationship between productivity and AI adoption at the economy-wide level.” Some workers report AI is actually making them less productive.
The Week’s Numbers
Total tech layoffs in 2026 have now reached 59,000, affecting 171 companies at a pace of 736 workers per day.
Recent announcements beyond Epic:
- CBS News: Cut 6% of workforce on March 20, shuttering its nearly century-old radio division
- Ingka Group (IKEA): 800 office roles cut on March 19 across 32 markets
- Crypto.com: 180 jobs (12% of workforce) on March 19, explicitly citing AI
The pattern from Q1 holds: Amazon leads absolute numbers (approximately 16,000), followed by Meta’s planned 15,000+ cuts. Block’s 4,000-person layoff in early March remains the largest single AI-attributed workforce reduction in history.
Where the Jobs Are: Data Centers and AI Infrastructure
While software roles shrink, physical AI infrastructure is creating demand for different skills. Six-figure salaries are now common for trades workers building AI’s physical infrastructure.
According to Randstad’s latest report:
- Robotics technician demand: up 107%
- HVAC engineer demand: up 67%
- Construction roles: up 30%
Data centers need electricians, cooling specialists, and technicians who can maintain AI hardware. CNBC reports these roles often pay six figures without requiring a college degree.
AI Roles: Premium Pay, Talent Shortage
For those with AI skills, compensation data from Ravio shows persistent salary premiums:
- AI/ML hiring grew 88% year-over-year
- AI/ML Engineer represents 45% of all AI job titles
- 12% salary premium for Professional-level AI roles
- Senior AI engineers baseline at $160,000+
The supply problem remains acute. Only 205 AI PhDs were awarded in the US in 2022, and over half of AI master’s and doctoral graduates are non-citizens who may not stay.
What This Means
The CFO survey data is the most interesting development this week. Companies are cutting jobs based on what they expect AI to do, not what it’s actually doing. That’s a gamble—if AI productivity gains don’t materialize, we’ll see companies that cut too aggressively scrambling to rehire.
For workers, the message is mixed. Entry-level and generalist roles continue evaporating. But anyone who can work on AI systems—whether coding them or maintaining the physical infrastructure—is in demand.
The question is whether the productivity gains will eventually arrive, validating this year’s aggressive cuts. Or whether we’re watching a rerun of the 1980s computer paradox, where the technology’s impact took far longer to materialize than early enthusiasm suggested.
Either way, 59,000 people have already lost their jobs in 2026. That number isn’t theoretical.