this post was submitted on 03 Dec 2025
557 points (98.1% liked)
Technology
81118 readers
4697 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
But the profit absolutely can materialize because it is useful.
Right now the problem is hardware / data center costs, but those can come down at a per user level.
They just need to make it useful enough within those cost constants which is 100% without a doubt possible, it's just a matter of can they do it before they run out of money.
Edit: for example, nvidia giving OpenAI hardware for ownership helps bring down their costs, which gives them a longer runway to find that sweet spot.
It is unlikely to turn a profit because the returns need to be greater than the investment for there to be any profit. The trends show that very few want to pay for this service. I mean, why would you pay for something that's the equivalent of asking someone online or in person for free or very little cost by comparison?
Furthermore, it's a corporation that steals from you and doesn't want to be held accountable for anything. For example, the chat bot suicides and the fact that their business model would fall over if they actually had to pay for the data that they use to train their models.
The whole thing is extremely inefficient and makes us more dumb via atrophy. Why would anyone want to third party their thinking process? It's like thinking everyone wants mobility scooters.
These companies have BILLIONS in revenue and millions of customers, and you're saying very few want to pay...
The money is there, they just need to optimize the LLMs to run more efficiently (this is continually progressing), and the hardware side work on reducing hardware costs as well (including electricity usage / heat generation). If OpenAI can build a datacenter that re-uses all it's heat for example to heat a hospital nearby, that's another step towards reaching profitability.
I'm not saying this is an easy problem to solve, but you're making it sound no one wants it and they can never do it.
Yep, I am. Just follow the money. Here's an example:
https://www.theregister.com/2025/10/29/microsoft_earnings_q1_26_openai_loss/
... That's all in your head, mate. I never said that nor did I imply it.
What I am implying is that the uptake is so small compared to the investment that it is unlikely to turn a profit.
😐
I've worked in the building industry for over 20 years. This is simply not feasible both from a material standpoint and physics standpoint.
I know it's an example, but this kind of rhetoric is exactly the kind of wishful thinking that I see in so many people who want LLMs to be a main staple of our everyday lives. Scratch the surface and it's all just fantasy.
You > the trends show that very few want to pay for this service.
Me > These companies have BILLIONS in revenue and millions of customers, and you’re saying very few want to pay
Me > ... but you’re making it sound no one wants it
You > … That’s all in your head, mate. I never said that nor did I imply it.
Pretty sure it's not all in my head.
The heat example was just one small example of things these large data centers (not just AI ones) can do to help lower costs, and they are a real thing that are being considered. It's not a solution to their power hungry needs, but it is a small step forward on how we can do things better.
https://www.bbc.com/news/articles/cew4080092eo
Edit: Another that is in use: https://www.itbrew.com/stories/2024/07/17/inside-the-data-center-that-heats-up-a-hospital-in-vienna-austria
And how, pray tell, will doing all of that return a profit?
I'm from Australia, so I can only speak to the Australian climate and industry. I can confidently say that the model shown in Vienna is not feasible in our country. We simply don't have much use for excess heat and we are highly susceptible to droughts. DCs use a lot of water to cool down and having these all over the country for private enterprise is bonkers. So, that's instantly a market that isn't profitable. Furthermore, it's not feasible to build a pipe and re-route the heat across large distances with minimal heat loss.
However, even when or if they implement this throughout all of Austria, it won't return a profit (which is what I thought your attachment was here, not the feasibility. We are talking about profitability, right?). This project cost $3.5m Euro and partially funded by tax. It's not a great example of profitability but a good example of sustainability measures.
Also, reading comprehension assistance: not feasible != Impossible.
Australia isn't the greatest spot to run a data centre in general in terms of heat, but I do understand the need for sovereign data centres, so this obviously can't work everywhere.
What makes you think $3.5 million can't be profitable? A mid sized hospitals heating bill can get into the many hundreds of thousands or into the millions even. Especially if it's in a colder environment. A 5-6 year payback on that wouldn't be terrible and would be worth an upfront investment. Even a 10 year payback isn't terrible.
These colder locations are the ideal locations for the data centres in the first place because they generally want a cooler climate to begin with, so they will gravitate to them when possible.
Edit: And if you build a data centre with this ability to recoup heat, you could start building further commercial things in the area and keep the heat redistribution very close. You don't need to travel very long distances. You do need to put some thought into where they go through and whats around or will be built around.
Ok. We're deviating off the point of LLM profitability here and have driven this conversation off into the weeds. So I'll make this one last comment, and then I'm done. This debate has been interesting but exhausting.
Final counterpoints:
LLMs have long since gone beyond the scope of interesting science project to something driven by pure parasitic greed.