As AI spending surges, computing costs at some companies are starting to surpass employee payroll, Axios reported on April 26 local time.
Bryan Catanzaro (브라이언 카탄자로), vice president of applied deep learning at Nvidia, said, "For our team, computing costs far exceed employee costs."
According to a report by The Information, Uber's chief technology officer also said that rising token costs used up the company's full 2026 AI budget in four months.
Amos Bar-Joseph (아모스 바-조셉), chief executive of Swan AI, said, "We are scaling with AI rather than hiring more people," drawing attention by sharing an Anthropic bill.
According to Gartner, global IT spending in 2026 is expected to reach $631 billion, up 13.5 percent from a year earlier. It is being driven by continued growth across AI infrastructure, software and cloud services. Proving results is needed to justify large-scale AI spending. Pressure is growing to show productivity gains or clear returns on investment.
AI pricing is also a variable. An OpenAI investor said Codex has an advantage in price competition because it uses tokens more efficiently than Claude Code, lowering usage costs.
Chi-gyu Hwang
delight@d-today.co.kr