Companies pushing AI automation are discovering that the cost of AI tools and compute power can exceed employee salaries, according to executives from Nvidia and Uber.
Bryan Catanzaro, Nvidia’s vice-president of applied deep learning, told Axios, “For my team, the cost of compute is far beyond the costs of the employees.” His remarks highlight growing concerns around token-based AI pricing as businesses increase their use of coding assistants and automation agents.
Most consumer AI assistants operate on subscription plans, but enterprise AI spending is often tied to token usage. Costs rise as companies run continuous AI-driven coding, automation and scheduled tasks.
Uber CTO Praveen Naga told The Information that the company’s AI spending estimates have already increased sharply. He said he had “[gone] back to the drawing board because the budget [he] thought [he] would need is blown away already.”
The issue is not limited to large firms. Swan AI’s Amos Bar-Joseph, in a LinkedIn post, shared that his four-person team received a $113,000 bill from Anthropic, the company behind Claude AI. That translates to roughly $28,000 per employee for a month.
The rising costs have triggered debate over whether AI is truly cheaper than human labour. A 2024 MIT study cited in the article found that humans performed the work more efficiently in 77% of cases.
Despite that, companies continue to invest heavily in automation. Uber’s Naga said AI agents now write 11% of the company’s live code updates. He added, “the vision for [him] as a CTO is to transform from software engineering to [AI] agent software engineering.”
Nvidia CEO Jensen Huang also reportedly believes engineers should increase their AI usage. According to the article, he wants engineers earning $500,000 annually to spend at least $250,000 a year on AI tokens.
Businesses are now weighing whether AI expenses are temporary investments before large-scale automation stabilises, or whether token costs will remain a permanent addition to salary bills.
At the same time, companies without clear AI strategies risk heavy losses. The article notes that many firms rushing into AI adoption without proper planning have struggled to generate returns.
Even so, layoffs linked to AI adoption are expected to continue as companies test the economics of automation and long-term workforce reduction.
