The funny thing about people who say it’s not a bubble because AI has value is that the asset category having value doesn’t prevent valuation bubbles from forming.
Houses have value: you can live in them. Yet there was a housing bubble.
The internet has value: you can watch cat videos on it. Yet there was a dot com bubble.
Tulip bulbs have value: you can grow pretty flowers with them. Yet there was a tulip bulb bubble.
In my experience, whenever you start reading news stories asking if something is a bubble and quoting investment bankers say, “no, it’s not a bubble,” well, usually it’s a bubble.
NVIDIA really out here selling shovels in the gold Rush
Nvidia are very smart in that regard, ethics aside. Very early on they decided that selling cards to gamers will not give them the infinite growth everyone so desperately desire, so they started looking for what does, and they were consistent at it ever since. Every tech bubble of the recent history is powered by Nvidia cards. How much they contributed to the hype (and damage) is not entirely clear, but that’s not zero for sure
They lucked into it. They made their cards for gamers, and various groups, AI researchers, bitcoin miners and others, discovered that they those gamer GPUs were really good for other tasks too. I think it took a while before Nvidia started making specialised cards for those purposes.
I can’t really blame them for serving that market that they just lucked into. I can and will blame them for their terrible Linux support.
If it’s just luck why isn’t AMD rolling in it with their cards?
Oh believe me, it wasn’t just luck. They have special labs full of people who’s whole job is to find another unexplored niches that can buy their cards. And they only make specific single purpose cards only when the market is mature enough to justify the spending, which is also smart.
They made gpus long before the gold rush and will not stop after. The usefulness of tensor cores will not dwindle with any market correction. Even before ai boom they were valued astronomically out of reality. Not a single stock is tied to actual selling or owning of anything anymore
Just like shovels existed before the gold rush and will exist after humanity’s death. But we have a saying for a reason
Hold up everyone. It’s not a bubble.
“So it is true that valuations are high but, in our view, generally not at levels that are as high as are typically seen at the height of a financial bubble,” said Goldman Sachs strategist Peter Oppenheimer.
He’s from GOLDMAN SACHS LOLOLOLO I THINK THEY WOULD RECOGNIZE A BUBBLE LOL ah fuck me our economy is gonna splode
Goldman Sachs also though NINA mortgages were a good idea, and they also thought it was a good idea to bundle bad mortgages in with good mortgages, and find a rater to mark them AAA investments.
And then we saw how that worked out.
yeah, how could this go wrong?
at least after the crash those houses could be lived in. these datacenters are made for one purpose, AI, and really would have to be completely gutted and refurbed for general purposes… fun.
I wonder what the people over at Bear Stearns think oh wait they gone.
this made me chortle into my cereal ty
I mean, what’s he gonna say?
Give me exit liquidity so I can buy the dip?
just short the companies that’ll be the most affected. probably nvidia is a good choice to short right about now.
I wish it were safer to make these sorts of sweeping gambles. Shorting Nvidia right now is like a pretty obvious bet but getting the timing right is the difference between generational wealth and a lifetime of poverty and debt.
right? I just figured the “it’s not a bubble, guuuuys” crowd could find someone a tiny bit more credible lol
There is definitely a bubble. But also what Nvidia is doing is smart. They have boatloads of cash. They are investing that cash in the companies that are using their products to create money making services. If one of them can create a killer app or viable service this will create demand for their products and they will have an ownership stake in it. Is this guaranteed or even likely? Probably not. We have reached the point where we were in 1996 where the chairman of the fed came out and said we are in a period of “irrational exuberance.” That bubble took four more years to pop. This one may end quicker, but it is impossible to tell when it will end or what will come out of it from where we sit today.
Unpopular opinion but this will not as bad as housing bubble and we’re way past bubbles actually popping in contemporary economy. Even China corrected for its massive ghost city housing bubble just recently and that was actually worse than ai tech overvaluation.
Can you explain how we’re beyond bubbles like I’m 5? Is it that there are gentler market corrections now?
Yes, contemporary economy and free markets are so imaginary now that cascading effects and bubble pops like 2008 are very unlikely. American stock market in particular is so far off reality (even before AI boom) that it’s basically a video game with no actual relevancy to true gross product. While China/Russia is a dictatorship with no representation of reality at all and can easily hide the burden of bad economic policies in the obedient peasant class.
So we have dictatorship with imaginary worlds vs “free markets” living in their own imaginary simulation. Economy is all made up now and cascades are basically impossible because that requires rationality.
Perfect explanation, also; since 07 thing where the hedge bros were not punished, there stopped existing any incentive to imagine any scenario where anyone lose any money due to bank runs
The problem isn’t the imaginary market, which I agree with the description. Its the leveraging of debt, to gamble in the market, which is what low interest rates enable.
And yes, our interest rates are VERY low still. I’m looking at some ARM packages right now, and their max lifetime interest rates are on par with what a typical mortgage was about a decade ago.
This turned out a little bit long. I wonder if anyone will bother reading it.
A lot of this so-called ‘bubble’ is based on capital expenditure in support of a technology that probably doesn’t have the capability AI company ceo’s claim, but does have fascinating, and in terms of how society is currently arranged possibly extremely harmful, potential.
I know what ai companies have done, and what they are likely to do, in the pursuit of profit is shit; I would say that is a capitalism and fascist billionaire issue, rather than a tech issue but ymmv.
And there is the energy consumption problem. I think ai ceo’s and tech broligarchs would privately say ‘compare the energy consumed by my datacentre to the energy consumed by the workers it has replaced and you will see it is fairly efficient…’ (I am saying what I expect they think, not what I agree with).
The concern that the economy currently has all of its eggs in the ai basket seems reasonable, but I see why capital is betting on it as big as it is. Any concerns regarding the economic disruption of an ai bubble popping is nothing compared to what could happen if 50%+ white collar workers are made redundant. We saw the number of essential workers needed per 1 million people during covid: it wasn’t many. Most jobs exist because the people exist to do them, corralled into the pyramidal structure of capitalism, where money trickles upwards. AI might push us into an era where the people exist but the jobs do not.
Anyway, I see this ‘bubble’ as being like the dotcom bubble, which didn’t kill the web when it popped. The gpu’s this capital expenditure has paid for are going to continue to be used, even as this economic period shakes itself out. They aren’t just going to evaporate. It isn’t like worthless debt being packaged up and resold without a chance of it being recouped, even if the prospect of what can be achieved with AI is currently over-valued.
Comparing to the dotcom bubble is what finally made it make sense in my brain. Though I know the toll it took on that sector’s workers and I don’t envy those in fields that are going to be affected the same way.
I’ve been saying the same thing.
The 2008 housing bubble was predicated on cheap lending. It was all debt. It was massive amounts of toxic debt sold around Wall Street, like using Trump Coin or counterfeit cash used to buy a house.
The vast majority of what’s happening here is not debt. Sure, some, but very little. Even the OpenAI AMD stock swap thing is swapping a gamble on stocks worth real money, not debt.
IMO the first sub-bubble to pop will be all the time and effort wasted on “Startups” that are nothing more than a couple people acting as a wrapper for an AI agent. That’s not really going to impact the economy too much on its face, but suddenly a lot of people are going to go from being “entrepreneurs” to being truly unemployed.
Edit: Also, just saw this gem, and THIS is how you get a supercharged 2008 repeat, bank deregulation and $2.6 trillion in lending. Which is exactly how we got to 2008’s subprime lending.
The vast majority of what’s happening here is not debt.
Most of what is going on in the AI sector is most certainly debt leveraged. Like, I’m looking at the books for several companies deep into AI.
I mean, how much profit is OpenAI turning right now?
I’m looking at the books for several companies
Well with all that proprietary information, please do enlighten us with specifics. Who has loans, and how much? From which banks?
Idk if ghost city thing was a bubble tho.
China used planned infrastructure and bunch of confused journalists in US were like “what kind of government plans for housing of their citizens”
It was a textbook bubble. They made and gambled on theoretical apartments where nobody involved had any intention of living there or any responsibility or connection to the underlying structure, to the point where building cardboard skyskrapers became a business… the is no point in denying it. Capital housing investment is a plague on humanity.
There are always bad actors in the system (see: hedge funds). But bubble? It can be argued that Ordos (the ghost city) was build too early, but it’s filling in nicely. From 30k in 2009 to 2.000.000+ in 2020.
https://en.wikipedia.org/wiki/Ordos_City
On the other hand noone ever build a damn whole modern city before for the people, so I’m not surprised they jumped the gun.
When Investors own houses only because it will appreciate in value from time only, that is a fundamentally flawed system because in reality houses decrease in value as they get older. It creates an environment where everyone involved is a bad actor
I’m sorry, I’m a little bit lost. I do agree that investment in owning rentals should be forbidden (and if city needs rental units they should be owned by the city).
I do not agree that “ghost cities” were built for speculative purposes. Speculants were buying them like crazy, yes, but the actual need for housing in regions planned (expected?) to undergo urbanization is real and the buildings were fulfilling that purpose.
Let me guess… is it because the people’s republic is flawless in every concievable way? I saw some footage from someone that was there, not a news agency. I won’t pretend that couldn’t be wrong, but I am known to occasionally enjoy Occam’s razor.
Yeah, my server end with ml so I must be a tankie, makes sense.
It cannot be that I genuinely appreciate long term vision policies. I must be a tankie. And you’re right, those cities were not needed, and planned urbanization must be planned only 2 years ahead because everything else is speculative bubble made for speculation.
I mean even if it was planned the amount of excess given falling birthrates, doesn’t check out either.
Ah yes “the stoopit west har har” propaganda lol
I was mostly going for “modern journalism is is sad and biased towards clickbait” ngl. Especially now they have AI edited articles.
I see “gold rush” the company selling shovels is making out like a bandit, everyone else is make a profit on the previous gen but requires a 10x cost increase for the next gen. And thus 10x more shovels… As soon as 10x more shovels stops giving 10x+ improvements this is the wrong investment.
Hints are we already reached this point.
Some AI companies will pivot and improve in other ways with more linear costs/results… The ones hoping the line continues to the moon… I think they overshot… I just don’t know when it will fall back…
I don’t think it’s a bubble, first there is absolutely zero comparison to the housing bubble, which was a financial problem that caused housing prices to inflate, while the inherent value of housing stayed the same. This alleged AI bubble is mostly driven by companies that have lots of money, so it is not credit based, and there are underlying products that actually have increasing value.
The better comparison would be the dot com bubble, which was dominated by companies that didn’t even have a product and didn’t make any money. The frenzy is similar, but the fundamentals are different.
AI investments may cool down because obviously there is a frantic race in an attempt to get ahead.
But the reason I don’t think the AI bubble will burst is because it is driven by companies that actually make money.
They may lose money investing too heavily in this, but the most companies investing in this can afford it.I think the most AI bubbly company isn’t even in the diagram, because that is Tesla. Tesla might actually go down, because Musk is insane.
But in general if it is a bubble, it is a very very long one, Nvidia value has been exploding since 2016 based on their AI product dominance. If this is a bubble, I think it will go down in history as the longest living bubble ever.
Is the market frantic? Yes absolutely.
Is the value of some AI companies extremely high? Yes absolutely.
Is it a bubble that will burst? No if it’s a bubble, this one will be more like deflating to a less frantic level, because ALL the main players have the money to weather losses.
And the main AI companies have actual products that make money for them rolled out already. So it is not like the dot com bubble.And the main AI companies have actual products that make money for them rolled out already. So it is not like the dot com bubble
Citation needed.
the only company making a profit is nvidia. everyone else is losing.
Not true, plenty of AI companies are in the green like Midjourney. Not everyone is hermorhagging money like OpenAi to curb out competition.
I don’t think the AI bubble will burst is because it is driven by companies that actually make money.
Last I looked, the big AI companies are all hemorrhaging money.
I think the biggest difference between this bubble and the ones that pop are whether the valuations were built by debt. In this case - no. So when their products turn out to be less useful than they claim, it will devaluate. But the debt issued to build the bubble wont go through a sudden correction that is amplified and causes an even bigger collapse like in 2008 or the dotcom bubble.
In bicycle repair terms it is called a slow leak.
Well argued. Also, even if it is a bubble, it’s arguable that most technology innovations are preceded by necessary bubbles which are important for directing investment into emerging technologies. The railroad mania in the 19th century or the fibre optic rollout in the late 1990s, during the dot com boom, benefited humanity long after the froth and excitement subsidied.
Its not a bubble but most people here dont think for themselves. They dont even seem to understand the connection between what news is put out, which analysts they choose to give attention, and for what purpose.
Imagine living your life and just believing whatever someone says in the news just because he has the title of analyst. And never thinking about who profits from that specific guy being on the news at that specific time. Who picked that guy to say what he does and why? Its not random.
Being able to influence the market is key to making a lot of money. How do people think they influence the market? This is how they do it. How else?
Sometimes they probably lose money too, specially when orange man opens his mouth and says something very stupid, like last Friday. But then they position themselves for the coming uptrend and make their money back, maybe even more then they had before, since they have giant pockets.
Its not a bubble but most people here dont think for themselves.
But the ones who believe the AI hype think for themselves. Right.
If people think it’s a bubble, then it’s a bubble! (Self-fulfilling prophecy.) Google Trends is a decent gauge of public sentiment. That said, the fundamentals are pretty flawed too.
Especially your first paragraph is probably spot on. Short attention span.
If not for the banks investing hevily into it, i’d not be all that worried.
Every company in that list could shrink by half and we’d all be at worst back to covid times. Sure unemployment would suck, but do we REALLY need microsoft and NVidia to be as huge as they are?
It seems like the wealthy propping up their own bubble.
Well, they now control all the money, so they can decide all the value.
All ai companies should direct all resources to medical research. I mean we would have to do without ai slop summaries for search engines and ai slop images. Well on second thought I guess slop is worth the human cost so let’s keep it as it is. I bet I get my wish.
The housing bubble encompassed a metric ton of banks and companies that bought and sold shares of subprime mortgages in the billions of dollars and when everyone stopped paying and started defaulting, that caused a entire economic collapse.
Now unless someone can point me to an analysis where we have some tangible proof that banks and tons of companies are invested, not just using, AI, it seems to me the fall out would be limited to tech companies, which yeah would involve some job losses but nothing on the scale of the housing or dotcom bubble.
Now if you’re referring to rich jackasses who are all in and banking on AI taking our jerbs? Sure that bubble will hurt them but they’re not driving forces in the economy, just politics, which I guess could cause a economic crash if they get your idiot politicians more scared of them than the people with France on their minds.
This only got downvotes in another thread. There is far worse that can happen than an AI bubble.
People get distracted over the fate that the pure speculative frenzy could be an AI bubble, and the harm to the hapless speculators and banksters could have a minor impact on the rest of the economy.
Reality is far worse than an AI bubble. It is a US mission for a fossil fueled powered Skynet for Israel that is too big to fail. Bubble in AI investments becomes unlikely, but total destruction of rest of US economy/prosperity becomes assured when the “plebs able to eat in America bubble” bursts is a sacrifice that a fossil fueled powered Skynet for Israel is willing to make.
If Americans are still able to afford to eat, then China or Iran wins.
What
I think they’re claiming TPTB are gonna try and tie dinosaur fuel and vapoware tech as “a matter of security” and constantly bail them out to the detriment of the US economy as a whole
OpenAI alone has 20gw of datacenter/gpu commitments. 20x entire current US corporate deployments. Sure as fuck, sweet US military/NSA overpayment for datacenter time is coming. Not sure AGI is needed to decide (family guy meme) “if it’s brown, flush it down” or “sink any boat you find in Carribean sea”. LLMs today would do pretty well at “make up a Department of War memo for backstories justifying all the people we killed today.”
That wasn’t an invitation to talk at me
please be more specific in what you don’t understand. I guess that…
fossil fueled powered Skynet for Israel
US government needs AGI for US military supremacy. That is Skynet (in Terminator movies, this is the military program to install AI in all the computers, and then AI chooses genocide nuclear launch). It is for Israel’s benefit, because that is who owns US government. That it be fossil fuel powered, serves another key US oligarchy. Skynet for disinformation/sediction detection purposes just as much of a threat than its use for nuclear genocide.
Regardless of whether datacenters will make money solving business and individual problems or boosting productivity, the US will keep investing in order to get Skynet. You can be correct that “frontier datacenter LLM models” will not make money, but still lose on financial bets validating that idea. Instead of an AI bubble bursting, even more money chasing Skynet will come with Austerity for rest of population. The “valuation bubble” only pops when investor money flowing in dries up. It may only dry up after the collapse of the US.
It’s the part where you say that Jews want skynet for world supremacy.
That’s the “what?” part.
Zionism controls US and its media. Nearly all politicians in US are Zionists even if just a handful are Jews. I’m not the one who said all jews are genocidal zionazi supporters. You have no right to try denying subhuman demonic evil zionazi supremacist control over the US by pretending, passive-agressive supremacistly, me or anyone else is making a remark about Jews.
you say that like skynet is a bad thing?
Anyone more knowledgeable care to help me understand where Anthropic is on this graphic/clusterjam?
Sell your stocks, I will buy. :)
Timing is a fools game for sure. Bubble could pop next month, next year, or even later.
If you’re old, make sure you have a good percent in bonds. If you’re young, make sure you have 6-12 months saved in case of layoffs and keep saving - market will look completely different in 20-30 years anyways so it’s not worth worrying about.
It’s objectively a bad thing when a country’s entire economy is being propped up by seven companies and the vast majority of consumer spending is concentrated in the top 1%.
Specially when those companies are valued in TRILLIONS. Nothing is worth trillions, somehow these surreal numbers have been accepted as hard fact.
Nothing is worth trillions,
There is things worth trillions. Like full countries, and the largest pension funds and social security funds. Having a single company be comparable to those massive collections of people is insane, and it’s because they think it can replace workers–when it can’t, not yet, and not fo a long time
Production must equal Consumption plus investment.
An excess of production leads to companies closing down.
Reducing consumption (via getting rid of workers, reducing wages, etc.) will lead to an imbalance that must correct itself.
This can be forestalled by private debt, government debt or exporting the surplus but this is unsustainable.
Evaluations of everything is crazy. Net worth of celebrities with make up lines in particular is crazy. Look how many celebs are worth a billion dollars. To be worth that much, they should be selling at least $50 millions a year of product with no prediction of winding down.
The most optimistic take I’ve seen: AI is a drain on the entire economy that sucks up all investment and this is why the rest of the economy is basically in a recession. Once the bubble pops, investors will flood back into the real economy and correct the problem.
I’m not optimistic.
Can the AI bubble please suck up all the housing investment?
The way to make a big dent in that is to tax unused housing, with peogressivwly increasing amounts as they continue unoccupied. And limit or outright deny ownership by companies and investment firms.
We have more than enough housing for everyone, but a large portion of it sits unused. In many cases only because no one will/can pay what some of these companies are demanding monthly for them.
I’ll play devil’s advocate here: agreed that the rest of the (US) economy seems to be slowing or shrinking but remains buoyed by AI / Mag 7 stocks. That said, a lot of the investment reflected above is in data centers and hardware (Nvidia, Coreweave, Oracle, Microsoft).
The bubble pop will hinge on whether there is value in this data center buildup beyond AI. Unless everyone starts paying fistfulls of cash for AI chat, these companies may be able to find another use for all that compute and avoid a total crash. That could be a target for all that investment you mention.
The hardware is specialized for chatbots, it’s not just something they can plug-and-play for other use cases. That means using it for other computing tasks is even less efficient per kWh and per litre of water, which will make it hard to justify the resource requirements.
Surely some of this hardware can find new life, but assets will be stranded.
Hahaha this is fine, I am fine with this, what could possibly go wrong
Basically Dutch Disease.
I feel money itself is our new Dutch disease. We live and die according to the flux of money in the global economy/stock markets…
Are there any theories like that out there? Because money start to no longer function correctly IMO.
Yep it’s such a fragile situation
