AI BriefingNvidia Executive: AI Costs Outpace Labor Costs
Key point
An Nvidia executive said current AI compute costs are higher than labor costs.
Details
Bryan Catanzaro said that, based on his team, compute costs far exceed employee costs.
Contrary to the narrative that AI immediately cuts costs by replacing workers, a 2024 MIT CSAIL study found that for vision-critical tasks, AI automation was economically viable in only 23% of cases, while for the remaining 77%, humans were cheaper.
The cost burden is also growing across the industry.
- Morgan Stanley estimated that AI-related capital expenditure has reached $740 billion so far this year.
- Layoffs.fyi tallied over 92,000 tech industry layoffs in 2026.
- The US Federal Reserve stated that as of late 2025, 18% of US companies had adopted AI tools.
However, there are also signs of a turning point. Gartner projected that by 2030, inference costs for 1-trillion-parameter LLMs could fall by more than 90% compared to 2025, and the industry expects AI to become cheaper and more predictable as pricing shifts from subscription-based to usage-based models.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Stanford and NVIDIA Release 'Contrastive Language Model' CLM with 9x Faster Inference