AI Computing Costs Outstrip Salaries, Challenging Corporate Returns and Cost Controls

Companies initially viewed generative AI as a way to reduce labor costs, but soaring spending on model training, inference and data centers has put returns on investment at the center of boardroom discussions. An MIT study estimated that only 23% of wages for tasks that could be performed using computer vision would be economical to automate, suggesting that AI is currently better suited to augmenting employees.

On April 26, 2026, Axios quoted Bryan Catanzaro, NVIDIA’s vice president of applied deep learning research, as saying his team’s computing costs were already far higher than employee salaries, though he did not disclose an amount. Gartner forecasts that worldwide IT spending will reach $6.31 trillion in 2026, up 13.5% year over year. Companies must therefore tighten controls on computing costs and validate returns on AI investments case by case.