AI Still Costs More Than Human Workers, Says Nvidia Executive Bryan Catanzaro

Artificial intelligence may be advancing rapidly, but it is not yet the cheaper alternative to human labour, according to Bryan Catanzaro, a senior executive at Nvidia. He believes companies expecting immediate savings by replacing employees with AI may be misunderstanding the current economics of the technology.

Speaking in an interview with Axios, Catanzaro said that for his own team, computing costs significantly exceed staff expenses. His comments challenge the growing belief that layoffs combined with AI adoption automatically improve profitability.

The remarks come as major tech firms such as Meta and Microsoft continue reducing headcount while simultaneously pouring billions into AI infrastructure.

Meta has reportedly planned workforce cuts of around 10%, impacting nearly 8,000 employees, while also freezing or removing thousands of open positions. Microsoft has also introduced one of its largest voluntary buyout programs in recent years.

At the same time, spending on AI is accelerating. According to Morgan Stanley estimates, major technology companies have already invested $740 billion in capital expenditure in 2026, representing a sharp increase from the previous year. Meanwhile, more than 92,000 layoffs have been recorded across the tech sector so far this year.

Studies suggest AI still lacks economic efficiency in many job categories. A 2024 study from Massachusetts Institute of Technology found AI automation made financial sense in only 23% of visual-task jobs, while human workers remained the lower-cost option in the majority of cases.

Experts say one of the biggest barriers is the high cost of computing power, data centres, and energy required to run large AI systems. Keith Lee of the Swiss Institute of Artificial Intelligence described the situation as temporary, predicting costs may fall sharply over the next few years as hardware and model efficiency improve.

However, affordability alone may not be enough. Analysts note that AI systems must also become more reliable, accurate, and easier to integrate into everyday business operations before they can truly replace human workers at scale.

For now, Catanzaro’s message is straightforward: replacing people with AI does not automatically mean lower costs, as human labour often remains the more practical and economical choice.

Source: Economic Times