Tech Firms and Large Employers just now realizing cost of compute is higher than paying human workers.
Nvidia VP Bryan Catanzaro told Axios that for his team, compute costs now run far beyond what his employees cost. Uber's CTO burned through his entire 2026 AI budget on coding tools alone — and that was by April. (Axios)
OpenClaw creator Peter Steinberger claimed that his team spent more than $1.3 million in token costs in just a single month. Because of this, it’s now apparent that using AI is more expensive than hiring people, especially since it offers only limited productivity gains at the moment.
Despite no clear evidence of AI improving productivity and no widespread data supporting the idea of AI displacing jobs, big tech firms have committed $740 billion in AI capital expenditures this year — a 69% jump from 2025. That spending has coincided with more than 92,000 tech layoffs in 2026 so far. (Fortune) So companies are simultaneously spending more on AI, laying people off, and discovering the math doesn't pencil out the way they projected.
2 replies (most recent on top)
Not only that but in June, Copilot will stop subsidizing people throwing it into agent mode. Which is what a lot of big enterprise companies are using, including us. I have my own subscription for personal work and use it pretty infrequently on the 20/mo plan. Under the new plan rolling out I would have spent over 100 for the same requests. And I've seen much more extreme comparisons like people who spent 40 and would have spent 5k.
These companies are removing people like crazy and are going to be in for a shock when they have to fork up the actual cost AND lost people with domain knowledge.
Just means they will layoff and offshore more to balance the books. Offshore + AI is too attractive of a proposition. However, this is when the lie of offshoring being the same quality will be exposed. Will it matter, no. What is considered good, will be lessened to make it appear to work.