hansolo

Awww...You don't even know the story?

https://en.wikipedia.org/wiki/The_Boy_Who_Cried_Wolf

"The Boy Who Cried Wolf" is one of Aesop's Fables, numbered 210 in the Perry Index. From it is derived the English idiom "to cry wolf", defined as "to give a false alarm" in Brewer's Dictionary of Phrase and Fable[2] and glossed by the Oxford English Dictionary as meaning to make false claims, with the result that subsequent true claims are disbelieved.

The only poor reasoning was the Boy, the main character, using bad info and emotion to raise false alarms until alarm fatigue brought harm to those around him.

If you are talking about GPUs being only a small share

My friend. I asked for a number.

You dodged and wrote a bunch of opinion.

No facts, no actual evidence presented. What good is this in convincing anyone of anything other than you're too deep in an emotional argument to ever see daylight again?

Shit, even someone else managed to copy and paste links to Boomer-targeted panic-y "news" articles. It's at least a tiny bit of effort to support their own opinions.

Hey, you want to see some real evidence that your concern is both overblown and potentially not even a thing at all? A pretty decent discussion on Hacker News about Dark GPUs, the term you should have been using, has plenty of industry-specific thoughts on the matter. The short version is that if Dark GPUs exist, it's Nvidia hoarding them, not companies waiting to use what they've bought. So even of the discussion about Dark GPUs, the custom aspect you're so fixated on doesn't seem to be real, either.

https://news.ycombinator.com/item?id=48446428

Not to mention your naive take that you think this is all actually about some very honest and straightforward hustle to stick GPUs in buildings.

Cute, but hey, let's get crazy, here.

Other factors to consider: This is a real estate investment scam to the core (see the article in the HN link). GPUs might hit EOL, but if used at a lower voltage, might be perfect for enterprise and free tier usage because those models are dogshit anyway. As access to compute returns to normal levels for Tier 1 users, if Dark GPUs don't end up used, they're a tax write-off; a massive, enormous, tax write-off for a company that's growing at stupid rates.

Hell, have you even considered that they don't even exist in the first place, and are one more slight of hand intended to push up GPU "scarcity" and inflate manufacturing and investment numbers? Nvidia's selling GPUs for 10 times above manufacturing costs. Sometimes 20x. Data center revenue is up 176% so far in 2026 because Nvidia is nearly a monopoly. And this administration won't touch them.

And now, to answer my own question, because you seem unable or unwilling: Most likely, zero percent. None, because Dark GPUs aren't even proven to exist. Let alone hyper-specific ones that you seem to have invented. Because demand for compute is so high that GPUs getting burned out with nothing to replace them is an actual problem constraining compute right now.

So until you have real evidence to back up your claims, I suggest you re-evaluate what you keep spouting off to people about, because it's not real. It's a hallucination, just like an AI would do.

In a $10 Trillion bubble, what percentage, exactly, are the GPUs you're taking about?

Is it 10%? Are there a Trillion dollars worth of GPUs you're worried about? Or less?

You're only cherry-picking some panic-mongering bias-confirming reporting here. Honestly - Yahoo? NBC? CNN? Why not Grandma's Facebook shares and the Babylon Bee?

Plus, did you even read the Forbes article? The opening paragraph is literally what I'm trying to tell you.

For example, one huge thing you're missing this is that the large companies of the Big 4 with frontier models aren't taking out loans for this. It's all the startups that are trying to draft on them that are using loans. THAT is a danger, that tens of thousands of idiots will sink a corner of the financial system from the sheer weight of defaulting loans because ChatGPT said their shit sandwich idea was "revelatory, and honestly, a great idea!"

But that's not going to tank OpenAI or Google. Google has cash money. OpenAI as a fair bit of cash money. Meta has cash money. Anthropic had cash and then Trump tried to sink them.... but they might be OK in the long run. They're doing great on the code side of things.

What you're also erroneously assuming points to a bubble points to how OpenAI, Google, Whatever Musk calls his stupid company this week, and Palantir all are doing to avoid classic bubble economics of huge loans to pay back. You don't pay back stocks. It's equity. That's cash trading hands, not loans that come due one day. Hell, Antropic is buying server time from Grok - that's real cash trading hands to perform a service. That's not a bubble, that's the kind of thing that prevents a bubble. Like, bro, do you even bubble?

You're not getting the full picture of the reasoning, or intentionally ignoring parts, I dunno.

• Large groups of people are historically bad at predicting financial markets. Very few people ever correctly predict a bubble ending, and considering that a large group of people are traumatized by 2008 and can read Wikipedia well enough to see the Dot Com bubble, they've erroneously put 2 and 2 together and think all large investments in tech will equate to a bubble. Regardless of the structure underlying it.

• Structural differences between Dot Com bubble and AI investments are numerous and extensive. Structurally, they're similar anecdotally at best. Yes, there are problematic parts. Data center demand will never be met by anything other than a few janky fly-by-night centers and ramshackle kludge-hosts in Serbia or Brazil where they're not regulated like the US or EU.

• The circular investment issue isn't just actual cash trading hands, it's assets and stock as well. In previous bubbles the majority of the bad investments were over-leveraged financing. Loans. There's actually very little in terms of loans going into these companies, which is a notable difference between this and literally every other bubble in history.

• I think the bubble will be 2 or 3 smaller bubbles that falter, but the mass of the overall industry will fail to full tip over because there's enough parts that can be scrapped and reapplied to other issues anyway, that demand won't ever evaporate as it did for $2 million URLs in the Dot Com bubble, or railway lines to nowhere in the 1840's.

• This does not ignore or assume no problems from layoffs and job displacement. That's a very real and huge threat, and AI will only enhance this problem by trying to claim it can manipulate and bilk poor people better than Google can.

Sure, some people heard about it in 2006. Not many, and even the people sounding alarms were called cranks up until the day after Lehman collapsed. However, right now a lot of social media is full of people basing the bubble's "any second now!" metrics on emotional AI-hate arguments, not real data points. I also thought we had crossed the line at some point, and that was 6+ months ago. Since then the Big 4 have been getting large contracts, which isn't exactly a sign of a hollow middle.

I'm sure we'll see in a couple years who was right, but I only see this as a fractional bubble where pieces fail individually, not the whole system as one.

The WorldCom fiber layouts were akin to the railroad bubble in the 1840s, in that those were pathways with nothing to use them. I can see the parallels here, but the difference is GPUs aren't nailed to the ground. They can be moved to demand, unlike railway lines and fiber lines.

GPUs process data. They don't spoil or expire. Sure, they'll lose value, but it's not like they stop being useful, even if highly specialized. Hell, even selling them second hand to China with an export waiver would be a way to recoup value. So already, the premise is flawed in that, specialized or not, China will use them. Or the EU or universities looking for a deal and building out their own local processing.

I haven’t heard this much bubble talk ever.

Exactly. And you know who's bad at predicting things? Large groups of people.

It's a fractional bubble at best. Unlike the Dot Com bubble, which invested in things like URLs that no one wanted and WorldCom's fiber layout, the buildouts and investments being done on spec aren't of the kind that are single-use. Data centers are always in demand, anyway. Storage and cloud compute are always in demand. Electricity is in demand. If AI flops tomorrow, other than 3 over-valued companies employing....a few hundred people, and idiot investors in those companies, the entire economy isn't yet entirely dependent on LLMs exactly because they suck at most of what they do anyway. Government contracts are stabilizing the whole industry anyway, something absent from the Dot Com era.

Edit: You know who IS a risk? All the dipshit startups that think putting a wrapper on a CustomGPT was a good business model, and took out loans for that rather than split equity. That's going to be your first indicator to look for, and it's wavering because, as it turns out, 99.999999% of those ideas are stupid.

History is the best teacher, and a detailed look shows these are only alike in that they are tech-related.

Ah yes, the absurdity of consistently inaccurate speculation being consistently inaccurate speculation.

Chicken Little vs. The Boy Who Cried Wolf.

Where's the flaw in the logic, again?

That movie was about 20 people copying one guy

Yeah, heard it all before, and I'm very familiar with the structural "curiosities" of the existing investment landscape.

Very few people correctly called the problems with 2007-2008. Not none, but few. And with soooooo many people mindlessly on the "it's a bubble!" bandwagon so early, a lot of accuracy and legitimacy is lost months or years beforehand for no other reason than why conspiracy theory people say "we'll get UFO disclosure this year!" Or "This year the Cubs/Arsenal/Red Sox will do it!" It's just the thing they say until one time they're right.

I'm not telling you it won't happen in a sense... But it's not going to happen how or when you think. IMO, you're looking at a partial stuttering effect maaaaaybe late winter like Q1 2027, and that's about it. There's to much alternate demand for everything LLM companies are already buying up to create a full and similar bubble like the Dot Com bubble.

You're talking about from buildup to crash, though. As if everyone just looking at literally any large investment and saying "it's a bubble!" is dong anything other than being a broken clock right twice a day.

I follow conspiracy theories extensively, and people have always predicted a huge, massive economic collapse next year - every year. On Art Bell, it was a constant, reiterated prediction every year from 1994 until 2013. It's only the ones that happened to say it in 2006 or 2007 that rode the credit of "actually predicting the 2008 crisis!" Even the ones saying it before the Dot Com bubble didn't get it right because their doomerism made all predictions "end of the world" level.