AI’s business model is starting to look less like a revolution and more like a $6 trillion IOU. Bain says existing AI services may cover just $1.8 trillion by 2031, leaving a $4.2 trillion hole to justify the data center frenzy. Apparently the killer app is still: »please invent enough revenue before the depreciation schedule catches us!!!«

The killer app is “you can’t build your own cloud because we’re hoarding all the compute”

Also many of the skills for building your private cloud have been forgotten or retired because Amazon and azure.

c/selfhosted

I’ve heard theories that I find convincing that the real plan is to transition us to a surveillance-based tiered economy in which the wealthy hoard control over the means of production and then spend generously to pay the largest segment of the economy – tech surveillance – to police the restless masses.

In this framing, the technology becomes profitable in the same way as healthcare: it’s essential. Once something is essential, its demand is unconstrained, and so its price-setting power isn’t based on the price customers consider a good value, it’s based on what fraction of their wealth and income they are able to hand over.

If the top 10% of Americans who participate in the investment economy conclude that ongoing real-time tracking of every American’s location and sentiments is necessary to prevent a revolution, paying to build and maintain datacenters is a must. It doesn’t matter what it costs.

This 10% then becomes the new real economy, and everyone in the 90% just become essentially livestock.

This is exactly what I’ve been saying. They aren’t building a product to sell, they are building a machine to replace and contain you. They don’t care about ROI, they care about the idea of complete control and they won’t stop chasing it until it kills them or they achieve their goals, the rest of us just get to die in the cogs along the way.

Even if AI takes off the way people hope it will, that does not mean that data center investments will pay out.

Local models exist and are a viable alternative (partially to substantially to completely) to sending every query to a centralized data center that is extracting profit from every token.

On top of that, everyone is now aware of the risks of tech dependency, and big players, the ones that would be sending their money to OpenAI or Anthropic, are the ones best able to declare their independence.

Also the vast majority of people don’t get any benefit from LLMs and won’t be willing or able to afford the actual cost to use them.

Hell, I would be surprised if the majority of people used AI for anything digital assistants or web searches were doing successfuly before they were replaced with AI bullshit.

Yes, certainly! It’s hilarious to me that people think the market broadly adopting a FREE product is somehow indicative of its economic viability. You just have to look at the journalism to see the chasm between use and profitability. And LLM’s, unlike almost every other form of modern software-based technology, does not scale. When Facebook doubles their user base, they double their revenue stream, at a fraction of a cost of exanding their costs. But when an LLM doubles its user base, it also doubles its costs.

AI is not a monopolizable product like Facebook, nor a locked-in key infrastructure ecosystem like Microsoft. It is a commodity, like toilet paper or oil or crackers. Anyone can enter the market to compete with you. No company will ever “own” AI–at least not this iteration of it.

Not even adopting really. Theyre shoehorning it in everywhere. When it’s a simple button in your chat app or search then of course people will use it. But they certainly aren’t going to pay for it.

That’s also ignoring the access to data question. Who knows how much the providers are logging. It wouldn’t surprise me if “new companies” pop up out of the blue with surprisingly competitive hardware and software sometime in the next decade.

In order to make $6 trillion by 2031, every human being on earth would have to spend at least $150 per year on average.

Maybe that doesn’t sound like a lot if you’re already paying a $100 a month for a Claude max subscription or whatever, but consider that the vast majority of people don’t pay for any AI at all, and have no interest in ever paying for it.

Also, $150 is a best-case scenario where revenue is 100% profit. In actual reality, the profit margins are much, much lower, and may not even exist at all!

every human being on earth would have to spend at least $150 per year on average.

6b humans would need to spend $1000/year on average. On top of phone/computer + internet/data services. Global average spending on food/year is $1500.

We’re up to 8.3 billion people now, fam.

6 trillion / 8.3 billion = ~750

And that number has to be spread over 5 years since the target date to raise the 6 trillion is 2031.

750 / 5 = 150

But, like I said, that’s still just a conservative estimate that doesn’t account for the fact that the AI firms have astronomical operating costs to cover.

The$6T number is per year revenue target in 2031. 6B is a round number, but facebook has 3b monthly active users, and google ai overview has 2.5b monthly active users. baidu (no google overlap) has 644m monthly active users. So at 3b internet connected users addressable or so, its $2000/user. MAUs double count work/personal/multiple/bot accounts.

Also, there’s about 1b desk/knowledge workers in the world. = $6000/job/year

Jesus… I totally misread the article. I thought it was 6 trillion total, not 6 trillion yearly revenue. That’s so, so much worse.

In case we’re not aware already, these companies are building faster and bigger than the market can keep up with. They’re expecting a tremendous fallout likely within the next several years. The companies that remain will dominate this market indefinitely.

The time bomb is society’s total prostration to the ruling class, of which tech journalism’s enabling vassalage is but a subset.

That’s changing, people are getting way fed up. The best years of the wealthy are behind them. The next two years are going to be violent chaos, as MAGA tries to increase their grip on power, and abolish elections, but that will be their downfall, and their oligarch slavemasters will go down with them.

That’s the problem and why they are desperate for AI. They don’t have a world changing idea like the smartphone or the Internet.

They have been lucky to have a new tech that becomes ubiquitous every 20 years or so - first radio, then TV, then Color TV, then Cable, then Home Video, then the digital age with CDs, cell phones, digital cameras, the Internet, then smartphones, which combined everything into one - phone, camera, video recorder, VCR, Internet, computer, calculator, calendar, etc., all rolled into one.

It’s kind of hard to beat the smart phone. The tablet has been a pretty good addition, and so has streaming, and now they think AI is the next big thing. It is becoming ubiquitous, and it won’t be a failed tech, but it won’t be everything that they are promising, either.

It’ll be like floppies, ironically. A niche specialized market

The MBAs think they can just pull an Internet and mobile device revolution out of thin air 💀💀💀

The Internet was a slow burn starting in the 1960s and gaining momentum through one of humanity’s most productive generations into the 80s and beyond, spurred on by personal computing. Mobile was a (retrospectively) natural continuation.

LLMs have been around for a very long time but they’re just forcing the issue with a product that’s not yet compelling enough to justify its enormous cost. It’s like waiting for the iPhone and the App Store to make smartphones ubiquitous.

Ehh I don’t think mobile was just a continuation of the Internet revolution, it was its own thing that happened to change the internet a lot.

I see it as an extension of the PC combined with the Internet and hardware advances. It certainly changed the Internet and our society. But I’ll leave it to historians to fight over!

Tactile screen was dreamed by TV. Mobiles covered that. On the technical standpoint, arm power consumption and having music, camera, gps, etc. all resonated with online/! digitalization of our selves. So, yes. I concur!

Mobile coincides with the idea of calculations and data being handled by 3rd party cloud servers instead of yourself.
That should be the time where the majority of performance shifted from being in peoples personal devices, to being in data centers owned by fewer large companies renting out access in various ways.

Not sure where else you wanna move the compute now for the next “innovation”. Even the bullshit about space datacenters isn’t really any different, it’s still compute owned by a company rented out.

Back to the devices (Apple with some local AI), but that’s not exactly revolutionary.

Vibe coding a new dot-com bubble

It’s fine. All of this is borrowed money and commitments. The reason prices on RAM and GPUs are through the roof is because all of the future capacity has been spoken for. But when the datacenters all that silicon is supposed go into don’t arrive, all of that booked demand evaporates. I think we’ll see a very nice market correction. Big tech will be fine, but stock prices will tank. The US AI labs will continue to exist, but the chinese labs will gut their competitive advantage. US put all their eggs into this stupid AI basket and China will be happy to give away the tech to spit in the US’s soup.

I take a market crash of apocalyptic proportions over a cloud dystopia any day.

You’re in luck because you’re going to have both!

Sure, I’ll just call the boys down at the lab and tell them to do that!

Those PhD students just need a grant

Once there is enough adoption, expect token costs to skyrocket and advertising enshitifaction to accelerate. We are still in the “first hit is free” stage of ai drug addiction.

So, that’s the best we’ll ever get?

The companies that replaced employees with AI are already entering the “find out” stage.

Once there is enough adoption, expect token costs to skyrocket and advertising enshitifaction to accelerate

Models on local hardware that give adequate performance and the availability of non-US models mean that path is doomed to failure. If they crank up the price then people will look elsewhere. While they might not get frontier-level performance out of the alternatives, there’s plenty of stuff out there that’s good enough for 1/4 to 1/10th of the cost. Or just the cost of power, if you’re willing to pay for some hardware and are operating on a small scale.

The only play the AI companies have got now is twofold :

• Regulatory capture, hence all the press releases of basic and frankly embarrassing sandbox failures and “we need to slow down” talk to try and get that rolling, and
• Banning of those nasty, evil, non-US models in the name of national security.

And then what? The rest of the world will just move on to the next most useful thing on the list. There will be ~330 million US citizens to extract wealth out of, to service trillions in data centre debt. That shit’s not gonna fly.

I’m sure they will keep hardware exorbitantly expensive for as long as possible to push ppl to the subscription model. I would love to run local models, but the cost of hardware versus performance is too high compare to cloud at this point in time. I might break even in 5-10 years, but the hardware may be out of date by then. I hope things will level out in a few years. And I agree the us gov will try to ban non-us models and hardware. They may even bail out the ai companies because they will be too big to fail. Also expect ai and robot companies to get massive military contracts.

I run Qwen 3.8 27B, which seems pretty comparable to Sonnet 5 for coding and other tasks. Qwen 3.6 was in the neighborhood of Sonnet 4.6. Qwen 4 27B is supposed to drop in the next couple months and we’ll see how it compares to Sonnet 5.5, which benchmarks at Fable / Mythos level.

If the trend keeps up, not sure how useful those data centers will be when a gaming PC is a few months behind the frontier.

Why do you think they’re doing ethics promises now? I have a feeling that all the third party models are going to fail the ethics tests and now that nvidia owns huggingface it’ll be very easy to ban them in the states.

Download your models now

Yep, I assume that’s the plan, though HuggingFace isn’t the only game in town. Hugging Bay, ModelScope, and I assume a horde of others will fill the void if HF goes downhill.

If they crank up the price then people will look elsewhere.

And, in last 2 weeks, the 2 leading US labs have reduced prices. Trend is going the opposite way.

A message to the industry:

Fuck yourselves, you traitorous, awful, slimy, scummy, pee brained, flee bitten, rabid, child raping, terrorist, ego driven, scumbag, sacks of monkey shit. Holy shit; hallelujah

I’m sure they will keep hardware exorbitantly expensive for as long as possible to push ppl to the subscription model. I would love to run local models, but the cost of hardware versus performance is too high compare to cloud at this point in time. I might break even in 5-10 years, but the hardware may be out of date by then. I hope things will level out in a few years. And I agree the us gov will try to ban non-us models and hardware. They may even bail out the ai companies because they will be too big to fail. Also expect ai and robot companies to get massive military contracts.

AI will save money, and make very little. The business will collapse, but the effects will be permanent in so far as they can. We’re stuck with some AI for good, unfortunately.

Just enough to keep people from going into withdrawals. A little AI, as a treat.

What the industry needs is to fail catastrophically and destroy the fortunes of as many billionaires as possible.

That would lead to an AI ice age. Surely not the worst development for Gaia and her children.

Innovation?

That’s just code for “Find a way to fuck consumers out of their money without actually giving them anything.”

Time to call the great innovator, Donald trump.

Claude, create an innovation which will dwarf what cloud and mobile unlocked. There we go. I’m a bit of an innovator.

My preparer randomly told me that he had invested money with a guy who was developing a “reactionless space drive”. He then mentioned that the guy had solved all the problems he was having with it by asking ChatGPT about it. I asked him if the guy had a working prototype yet and he said no but it was sure to work.

Time to find a new tax preparer.

a fool and his money will soon be parted.

You forgot „make it good“. You will never be a prompt engineer at this rate!

I said it years ago: it’s hard to improve on a tiny personal computer with millions of free apps and access to all the people and information in the world. The next step will be some sort of brain-computer interface that will let you access all the information instantly but it’s decades away.

Working on it.

💪😎✌️

you know where innovation comes from?

independent research.

you know what the oligopolies hate?

independent research.

pick one. you can’t have it both ways.

the financial logic behind the data center buildout simply doesn’t hold up.

need to change assumptions that it requires a decent end ROI, or good economic proposition, when analyzing logic.

1. Greater fool theory. Hype hype hype to cash out at IPO or buyout. Circular financing is your friend. Outright fraud of 5x more expensive data centers in space that are only useful for slow inference.

2. There’s nothing else for the US to spend on, and aspirationally seek economic competitiveness with China. A race to end US employment for higher oligarch profits, as a bribe to narcissist in chief, becomes somehow MAGA.

3. Skynet government surveillance, fascism, and geopolitical dominance. Government will buy all of the tokens or China wins. US fascism/oppression will increase to protect Israel first establishment power, or China wins. There can be no guardrails on US government, or AI bond villain, use of models, or China wins. Zero regulatory discussion exists on preventing the pursuit of Skynet, and “whoever wins AI, wins” is only/purely a Skynet goal, and it is policy unanimity. …blogspot.com/…/a-global-prosperity-manifesto.htm…

An irony in 2nd/3rd point is that industry is fully aware that about 90% of tokens in SEA area datacenters are consumed in China, and it is one of biggest areas of new datacenter developments. So some chain of control in the profit/revenue flow from token sales to China is a near term priority over skynet dominance over China.

I believe this is report all those articles are referring to bain.com/…/new-innovation-is-required-to-fund-ais…

“What the industry needs is a wave of innovation”

Narrator: They didn’t innovate

The only innovations made are new and unique ways to screw taxpayers and the general public

They definitely won’t with AI. It’s rotting skills.