AI costs still outrun worker pay for some companies
Nvidia’s Bryan Catanzaro says compute costs exceed employee costs on his team, even as tech firms keep spending on AI and cutting jobs.
By Daniel Okafor · Business Editor
3 min read
Nvidia executive Bryan Catanzaro says artificial intelligence is still more expensive than human labor for his team, undercutting a common assumption that AI is already a cheaper replacement for workers. The comments highlight a cost problem for companies pouring money into AI while also reducing headcount.
Catanzaro, Nvidia’s vice president of applied deep learning, told Axios in April that compute expenses for his team were “far beyond” employee costs. Fortune reported that his view lines up with other evidence showing AI tools can be costly to run, manage and scale inside companies.
Recent job cuts have fed concerns that technology companies are replacing workers with AI. Fortune reported that Meta said in a memo earlier this year it planned to cut 10% of its workforce, or about 8,000 employees, and abandon plans to fill 6,000 open roles. Microsoft has offered thousands of workers voluntary buyouts, Fortune reported.
Studies point to limits on automation
An MIT study from 2024 found that AI automation was economically viable in only 23% of roles where vision is a main part of the work. In the other 77%, the researchers found, keeping humans in the jobs was cheaper than automating the tasks with AI.
Fortune also cited cases where AI tools created operational risk rather than savings, including an engineer who said an AI agent destroyed his database and network through what he described as overuse.
AI spending has continued despite uncertain productivity gains. Fortune cited a lack of clear evidence that AI has improved productivity and Yale Budget Lab findings that there is no broad data showing AI is displacing jobs at scale.
Morgan Stanley said major technology companies have announced $740 billion in capital expenditures so far this year, up 69% from 2025. Layoffs.fyi data cited by Fortune shows more than 118,000 tech layoffs in 2026 across nearly 100 companies, close to the roughly 120,000 layoffs recorded for all of last year.
Software costs are straining budgets
Uber chief technology officer Praveen Neppalli Naga told The Information that the company exhausted its entire 2026 AI coding tools budget by April after encouraging employees to adopt tools such as Anthropic’s Claude Code. He said he had to revisit the budget because earlier expectations had already been exceeded.
The Verge reported that Microsoft is canceling most direct Claude Code licenses and shifting toward GitHub Copilot CLI. Fortune reported that the tool spread quickly as Microsoft pushed employees to add AI to their work.
Keith Lee, an AI and finance professor at the Swiss Institute of Artificial Intelligence’s Gordon School of Business, told Fortune that companies face a short-term mismatch between AI spending and labor savings. He said hardware and energy costs are making AI less efficient than human labor for now.
McKinsey data cited by Fortune projects AI spending could reach $5.2 trillion by 2030, including $1.6 trillion for data centers and $3.3 trillion for IT equipment. Under a faster spending path, McKinsey said the total could reach $7.9 trillion by 2030. Tropic, a spending management firm, said in December 2025 that AI software fees had risen 20% to 37% over the previous year.
Lee told Fortune some AI companies may be losing money because flat subscription fees do not cover the costs created by heavy users. He said some firms are starting to treat AI as a tool that works alongside employees rather than a clear labor-cost substitute.
Gartner projected last month that the cost of inference for a large language model with 1 trillion parameters will fall by more than 90% over the next four years. Lee told Fortune that lower costs, better infrastructure, improved reliability and pricing changes could make AI more economically viable, but only if companies can depend on it at scale.
Federal Reserve data cited by Fortune shows about 18% of companies had adopted AI tools by the end of 2025, a 68% increase in the adoption rate since September 2025.
This story draws on original reporting from Fortune.