- cross-posted to:
- programmer_humor@programming.dev
- cross-posted to:
- programmer_humor@programming.dev
https://www.axios.com/2026/04/26/ai-cost-human-workers Uber’s chief technology officer already blew through his full 2026 AI budget due to token costs, according to The Information.
Lol. Lmao even
Does AI cost more than humans primarily because of greed (i.e the AI companies demand a high profit margin now) or because of energy costs (i.e AI is so wasteful with energy, so polluting, that it costs more than human workers)
Given the ai companies are running at a loss, it’s fair to assume which of these is likely
This is a common myth, inference is not typically run at a loss, despite claims. It’s only a loss if you include staff and ongoing training costs. They could lock in their models now and be profitable if they wanted to.
Edit: I see the comment above has changed (or I misread initially) to say the companies are running at a loss rather than inference running at a loss. Yes, that’s extremely true. Now my comment doesn’t make any sense and is irrelevant so feel free to ignore my pedantry.
Yes, and let’s also not count all the investments in infrastructure because you know… like training and staff it’s not a real cost that’s essential to the business.
Anyways, you wouldn’t happened to have heard that from Anthropic or OpenAI?
Somehow we don’t have any actual indisputable numbers (I wonder why) but it is actually quite controversial and some of those who have done deep research on the subject are saying inference IS run at a loss and it might not get profitable ever.
https://www.ft.com/content/fce77ba4-6231-4920-9e99-693a6c38e7d5?syn-25a6b1a6=1
We do have numbers from comparably sized Chinese models.
Yes, every AI company is bleeding money, they’re not healthy in any way. But inference by itself is profitable, based on everything that we know.
Inference + amortizing the training costs is NOT profitable, which is what most people are talking about.
This is easily fixed by not releasing a slightly different version every month.