AI costs more than employee salaries at some companies: here’s why<\/figcaption><\/figure>\n<\/div>\nUber’s chief technology officer spent its entire 2026 AI budget before the year was properly underway, not on hardware or headcount, but only on token costs. This is an extreme case, but not an isolated one, and it points to ongoing change in corporate IT departments around the world.<\/p>\n
Nvidia implements deep learning, vice president Brian Catanzaro explains axios <\/i>The cost of computing for their team already exceeds the cost of employing people. The reversal where processing power costs more than the workers using it would have seemed impossible even two years ago. <\/p>\n
The preaching has also started. Swan AI CEO Amos Bar-Joseph posted publicly on LinkedIn about his Anthropic Bill going viral, calling it a badge of honor. “We’re building the first autonomous business scaling with intelligence, not with headcount,” he wrote. Whether this deal turns out to be a good one or not is a question that investors will ultimately have to answer.<\/p>\n
According to research firm Gartner, worldwide IT spending is projected to reach $6.31 trillion in 2026, a 13.5 percent increase from 2025. The growth is being driven by AI infrastructure, software, and cloud services, covering everything from large-scale model deployment to the cost of individual AI subscriptions at the enterprise level.<\/p>\n
That growth rate reflects real organizational appetite. But appetite and accountability are two different things, and the latter is arriving on time on quarterly earnings calls everywhere.<\/p>\n
“The tone is changing a little bit more on what is the real value of a worker, human or digital?” said Brad Owens, vice president of digital labor strategy at Assemble, a workforce orchestration firm. This is a question that boards and shareholders are starting to ask, along with productivity metrics and margin targets.<\/p>\n
The competition between these AI labs is increasing the complexity. According to an OpenAI investor speaking to Axios, Codex is considered more affordable with tokens than cloud code, which could impact enterprise spending if the cost burden increases further. Additionally, Anthropic has already changed its pricing structure due to increased demand.<\/p>\n<\/p><\/div>\n