Nvidia Exec: AI Compute Costs More Than Employee Salaries
Bryan Catanzaro, Nvidia's vice president of applied deep learning, told Axios in late April that his team's compute costs far exceed their employee salaries. This statement highlights a surprising economic reality in artificial intelligence: despite massive tech industry spending on AI infrastructure — $740 billion in 2026, up 69% from last year — AI remains more expensive than human labor in most real-world applications. A 2024 MIT study found automation cost-effective in only 23% of roles requiring vision, with humans remaining cheaper in 77%. Companies like Uber have already exhausted their entire 2026 AI budgets due to token costs. Analysts say this represents a short-term mismatch between AI's promise and current cost structure, though some predict prices will eventually fall. Meanwhile, Big Tech companies continue cutting jobs citing AI, even though productivity gains remain unclear and evidence of actual AI-driven displacement is limited.
Key facts
- Bryan Catanzaro, Nvidia VP of applied deep learning, stated in April 2026 that compute costs for his team far exceed employee salaries
- Big Tech companies announced $740 billion in AI capital expenditures in 2026, a 69% increase from 2025
- 2024 MIT study found AI automation economically viable in only 23% of vision-based roles, with 77% remaining cheaper for human workers
- Uber exhausted its entire 2026 AI budget by April due to token costs from intensive use of AI coding tools
- Tech sector announced 38,242 job cuts in May 2026 alone, with AI cited as the most common reason for layoffs across three consecutive months
How different outlets framed it
Left-leaning coverage: Left-leaning critics frame the scenario as tech companies using AI as cover for layoffs they wanted to make anyway. They highlight that firms are slashing workers while human labor remains cheaper, suggesting budget reallocation rather than genuine automation. Critics point to insufficient evidence of AI productivity gains and characterize company claims as 'AI washing' to avoid accountability for workforce reductions unrelated to technology necessity.
Right-leaning coverage: Right-leaning perspectives emphasize innovation and market efficiency. They argue Nvidia's statement validates that companies must invest in AI infrastructure now to remain competitive, even at short-term cost. The narrative suggests skepticism toward worker protection arguments, focusing instead on business logic: firms overhired during the pandemic and are now rightsizing while positioning for future growth through AI leadership.
Limits of this briefing
Sources generally lack data on actual productivity gains from current AI spending, making it difficult to determine whether AI is genuinely replacing workers or simply consuming reallocated budgets. Attribution of layoffs to AI varies widely by company and may conflate multiple causes including post-pandemic overhiring, economic pressure, and strategic pivots.
Sources
- Fortune: 'The cost of compute is far beyond the costs of the employees': Nvidia exec says right now AI is more expensive than paying human workers Centre
- Tom's Hardware: Nvidia exec says AI is more expensive than actual workers — yet some companies don't see the extra costs as a negative Centre
- CNN Business / Allison Morrow: Big Tech promised AI would disrupt labor — just not like this Left
- Tekedia: 'The cost of compute is far beyond the costs of the employees': Nvidia executive admits AI is more expensive than human workers Centre
- TechRadar / Industry Analysis: Nvidia says AI compute costs still exceed human labour expenses Centre