AI Still Costs More Than Human Workers, Says Nvidia Executive

AI Costs Continue to Outweigh Labor Savings

Although technology companies have announced workforce reductions and increased AI spending, some executives argue that artificial intelligence is not yet reducing overall costs.

Bryan Catanzaro, Nvidia’s vice president of applied deep learning, previously noted that computing expenses for his team exceed personnel costs. His comments highlight a broader challenge facing organizations that are rapidly adopting AI technologies.

Research from MIT published in 2024 reached a similar conclusion. The study found that AI automation was economically practical in only a minority of vision-related jobs, while human workers remained the lower-cost option in most cases.

Companies Continue Investing Billions

Even with ongoing concerns about costs, major technology firms continue to expand AI investments. Industry estimates indicate that Big Tech companies have committed hundreds of billions of dollars toward AI-related infrastructure and development this year.

Some organizations have already seen AI-related expenses exceed expectations. Uber's technology leadership acknowledged that adoption of AI coding tools quickly exhausted the company’s planned budget, while Microsoft has adjusted its approach to developer-focused AI tools after strong internal usage increased costs.

At the same time, layoffs across the technology sector have continued, with more than 118,000 job cuts reported across nearly 100 companies in 2026.

Why AI Remains Expensive

Experts point to several factors keeping AI costs elevated, including data center construction, specialized hardware, energy consumption, and ongoing operational expenses.

According to AI and finance professor Keith Lee, many providers also face challenges with subscription-based pricing models that may not fully cover the costs generated by heavy users.

As a result, some businesses are increasingly viewing AI as a tool that complements employees rather than replacing them entirely.

What Could Change in the Future?

Analysts expect AI costs to decline as hardware improves, infrastructure expands, and model efficiency increases. Industry forecasts suggest that the cost of running large AI models could fall significantly over the next few years.

Lee also expects providers to move toward usage-based pricing models, creating a closer relationship between consumption and cost.

However, lower prices alone may not be enough. AI systems will also need to become more dependable, reduce errors, and require less human oversight before companies can confidently rely on them at scale.

According to Lee, the long-term question is not simply whether AI becomes cheaper than human labor, but whether it can deliver predictable and reliable performance across business operations.