Bryan Catanzaro says AI compute costs exceed employee costs

4 articles · Updated · PC Gamer · May 8

The Nvidia vice-president for applied deep learning said his team’s computing bill is “far beyond” staff costs, underscoring the scale of spending behind AI development.
The remark highlights a tension in AI economics as companies promote automation to cut labour costs while facing heavy infrastructure and token expenses.
Nvidia has been a major beneficiary of the AI boom, while debate continues over whether the technology will replace jobs, create new ones or concentrate pricing power among a few model providers.

As AI's running costs eclipse human salaries, is the tech industry building a future it cannot afford to run?

AI promised to replace expensive jobs, so why is human labor still cheaper for the vast majority of tasks?

AI's thirst for electricity is threatening national power grids. What is the hidden environmental price of this technological revolution?