Companies are now discovering that AI more expensive than humans has become a surprising reality. Businesses are spending more on AI tools and computing power than on employee salaries, raising new questions about the real cost of automation.

Bryan Catanzaro, Nvidia’s vice president of applied deep learning, told Axios that compute costs for his team now far exceed employee costs. Axios also reported that Uber’s chief technology officer has already used the company’s full AI budget for 2026 because of soaring token costs.

So what changed? For years, automation promised to cut labor expenses. Now the math looks different.

Token Costs Become a Budget Issue

The rising cost of AI use is becoming a bigger concern as companies deploy tools for coding, research, support, and daily operations. Every query, every response, every generated image burns through tokens. Those tokens add up fast.

Amos Bar Joseph, chief executive of Swan AI, drew attention online after posting about his Anthropic bill. He said the company was building an autonomous business by scaling with intelligence instead of headcount. That strategy worked well until the bills arrived.

These examples show how AI expenses can increase quickly when companies rely heavily on paid models and token-based usage. In many cases, AI more expensive than humans has caught finance departments off guard.

Global IT Spending Keeps Rising

Worldwide IT spending is expected to reach $6.31 trillion in 2026, up 13.5 percent from 2025, according to Gartner. The research firm said strong demand for AI infrastructure and software drives this increase.

Gartner also expects data center systems spending to grow 55.8 percent in 2026. Software spending is forecast to rise 15.1 percent. These are not small adjustments. They represent a massive bet on AI.

Companies Need to Prove the Payoff

Higher spending may create pressure for companies to show clear returns from AI investments. Businesses with large IT budgets will likely need to prove that AI tools are improving productivity, reducing costs, or helping generate revenue.

That pressure may be even stronger for public companies that answer to shareholders each quarter. A CEO cannot simply say “we spent millions on AI” without showing results.

Brad Owens, vice president of digital labor strategy at Asymbl, told Axios that the conversation is shifting toward the real value of a worker, whether human or digital. If a digital worker costs more than a human worker, what exactly is the point?<h2>AI Labs Could Face Spending Pushback</h2>

Rising AI costs may also affect enterprise spending at major AI labs. Axios reported that an OpenAI investor believes the shift could benefit the company if Codex uses tokens more efficiently than Claude Code. The report also noted that Anthropic has changed pricing to respond to higher demand.

The larger issue is clear. As AI labs raise prices and usage grows, heavy AI spending could move from a sign of innovation to a financial risk. When AI more expensive than humans becomes the norm, companies will start making difficult choices about where to cut back.

For now, the automation dream has hit an unexpected wall. Efficiency no longer comes cheap.