Nvidia VP Says AI Compute Costs Now Exceed Employee Salaries

A senior Nvidia executive has admitted that infrastructure costs for his team now surpass payroll expenses, highlighting the high price tag of scaling artificial intelligence.

Nvidia Executive Questions AI Cost Efficiency

“For my team, the cost of compute is far beyond the costs of the employees,” Bryan Catanzaro, Nvidia’s vice president of applied deep learning, said in an interview late April.

His words carried weight because Nvidia’s chips power the generative AI boom, and the company now ranks as the world’s most valuable semiconductor firm. Catanzaro did not offer specific figures. But his observation exposed a growing tension. Companies are pouring billions into AI infrastructure while cutting headcount in the name of efficiency.

Uber Blows Through AI Budget in Four Months

The cost pressures Catanzaro described materialized quickly at Uber. The ride-hailing company burned through its entire planned AI budget for 2026 within four months, Chief Technology Officer Praveen Neppalli Naga said in mid-April.

“I’m back to the drawing board, because the budget I thought I would need is blown away already,” Naga said.

Uber gave engineers access to AI coding tools including Anthropic’s Claude Code and encouraged usage. The company built internal leaderboards ranking engineers by AI token consumption. By March, 95% of Uber’s engineers used AI tools monthly, and AI generated nearly 70% of all committed code.

Monthly API costs per engineer ranged from 500 to 2,000. One software engineer in Stockholm told The New York Times: “I probably spend more than my salary on Claude.”

MIT Study Finds Human Labor Cheaper in Most Vision-Based Jobs

A 2024 MIT study supported the economic concerns raised by Catanzaro and borne out by Uber’s experience. Researchers analyzed whether AI could replace human workers at a competitive cost and found that AI automation made financial sense in only 23% of jobs where vision was a primary task. In the remaining 77%, human labor remained cheaper.

No widespread data shows AI is displacing jobs or improving productivity at scale, according to the Yale Budget Lab. Despite this, Big Tech firms announced $740 billion in AI capital expenditures for 2026, a 69% increase from 2025, according to Morgan Stanley estimates.

“As a result, some firms are beginning to reevaluate AI not as a clear cost-saving substitute for labor, but as a complementary tool — at least until the cost structure stabilizes,” Keith Lee, an AI and finance professor at the Swiss Institute of Artificial Intelligence’s Gordon School of Business, shared.

Tech Layoffs Surge as AI Spending Explodes

Tech layoffs accelerated in 2026 even as the economics of AI remained unproven. According to Layoffs.fyi, more than 92,000 tech workers lost their jobs across nearly 100 companies in the first four months of 2026. The pace far outpaced 2025, which saw about 120,000 layoffs in total.

McKinsey projected total AI spending would hit 5.2 trillion by 2030, including 1.6 trillion from data center spending and $3.3 trillion from IT equipment. Fees for AI software increased 20% to 37% over the past year, according to spending management firm Tropic.

Lee predicted the cost of using AI would eventually drop. Performing inference for a large language model with 1 trillion parameters would plummet by more than 90% over the next four years, according to a report from analyst firm Gartner. For now, however, a human being often cost less than the AI tool used to do the job.

Chain Street’s Take

When the company selling shovels tells you the gold rush burns money faster than paying miners, the math deserves scrutiny. The MIT study’s 23% figure is the real story: AI makes financial sense in less than a quarter of vision-based jobs. The layoffs are not about efficiency. They are about narrative. The bill is coming due.