A $200 ChatGPT subscription could cost OpenAI $14,000 if you actually used it to its full potential (TechSpot, 6/14/26)

[highplainsdem]

https://www.techspot.com/news/112759-openai-anthropic-cant-afford-have-everyone-use-ai.html

-snip-

SemiAnalysis has calculated how big that gap really is. After testing subscription tiers from both OpenAI and Anthropic – running long-horizon coding and agentic tasks until weekly limits were exhausted – the firm found that the cost of theoretical maximum usage of these plans if priced at standard API rates far exceeds what users actually pay.

A $200 ChatGPT Pro 20x subscription could cost as much as $14,000 in API pricing if fully utilized. Anthropic's Claude Max 20x plan, also priced at $200 per month, has a comparable ceiling, with potential usage totaling roughly $8,000 in token costs.

Those figures help explain why utilization rates matter so much to the AI companies offering them. According to SemiAnalysis, Anthropic breaks even on Claude Pro and Claude Max 5x at around 20% utilization. OpenAI's margin is thinner. It begins losing money on ChatGPT Plus and ChatGPT Pro 5x once usage climbs above 11.4%.

The economics get tighter at the high end. Anthropic reaches zero gross margin at roughly 10% utilization on its top-tier plans, while OpenAI crosses into negative territory at just 5.7%. It doesn't take extreme use for these subscriptions to turn unprofitable.

-snip-

Ed Zitron has been pointing this out for a long time, but the article doesn't mention him even once.

[VBNMW_Realist]

1. My Take On Taking Apart AI

I hope people start using these models to try to learn more about the system. Not even to overload them in a financial DDoS-style attack where a bunch of people start trying to overload the model's usage to make it so it costs tons of money (which is actually technically legal). But having crowd-sourced projects to collect the data from these models.

And if these companies complain about their intellectual property, they are wrong for these two reasons. First, they "stole" other people's content. (I don't think it's really stealing in my opinion but still.) Second, the responses to the user are now or at least should be the property of the user, and the copyright and patent rights of these corporations should not override the users when they are doing their own thing

[highplainsdem]

2. The output from those genAI models is full of hallucinations.

And btw, the AI companies DID steal the world's intellectual property for training data. However much they try to excuse that theft now, there's plenty of evidence they knew what they were doing was unethical and illegal.

[VBNMW_Realist]

3. I agree.

The AI companies were using data illegally and that is not okay. What I was saying is that if they are allowed to, other people should be too (no double standards for the rich).

The thing about hallucinations is very interesting because they actually imagine things when they can't see what the dta is.