I think those IPOs are meant to bring in enough money to give the founders and managers golden parachutes before the whole system folds.
They don’t make any money - on the contrary - and they won’t miraculously start making it just because the IPO happened.
You’re thinking too long term. IPOs are the C suite cashing out.
Even the street corner hustler knows you give the first rock out for free.
They also know not to spend trillions making the product.
Hey, cocaine and baking soda ain’t cheap you know. Plus you got to have a microwave and shit. Overhead is crazy like that.
/s
Here’s the problem… He says AI was adopted beyond his expectations. Great.
But if somebody is using it at the current price point of super cheap or free, are they going to keep using it when it gets expensive?
You can make a basic chatbot run on a desktop PC, but nobody wants to pay for that. Once you get into things with useful generation and large context windows, or things like video generation, suddenly you need one or more $10,000+ pieces of hardware to run it. So the $10 a month you charge the user is basically an introductory price that doesn’t cover your hardware fees let alone the software engineers to build your AI.
Eventually, the bill comes due. Eventually, you have to look at your customers and how much machine time they use each month and how much your r&d costs and figure out what the actual cost to the customer has to be. And then the customer rethinks how useful the AI is or isn’t.
People will pay $10 a month for chat GPT to write their emails. Will they pay $100 a month?
What about the company that replaced all their software developers with AI. Suddenly the AI cost as much or more as the software developers. Only now the developers who understood the code base work for other companies.
There will be a fun correction when this happens.
Once you get into things with useful generation and large context windows, or things like video generation, suddenly you need one or more $10,000+ pieces of hardware to run it.
A Blackwell server with 72 GPUs costs about $3 million, plus requires 130 kW of power (about 3 residential homes’ max rated power through a residential 200A circuit box, for about $600-$1000/day in electricity cost).
You’re gonna need to sell a lot of $20/month subscriptions to get that paid for, assuming that the server is good for 5 years. If it’s only good for 3 years, the economics are basically impossible.
… assuming that the server is good for 5 years. If it’s only good for 3 years, the economics are basically impossible.
And surely the big AI companies wouldn’t decide that hardware rated for ~3 years should be amortized over 5-6 years in order to massively inflate their value on paper.
I’m sure Bernie Madoff’s fund had an ‘accounting puzzle’ too.
I love how the publications call it a ‘puzzle’ rather than what it so obviously is- a bunch of people throwing $billions at tech they don’t understand which has no serious road to profitability anytime soon.
If it’s only good for 3 years, the economics are basically impossible.
Also consider that as models improve, the newer frontier models that are doing actual useful work require significantly more compute and RAM than basic chat bots…
I am no AI fan, but I do not see this going away. I think if you can get enough people addicted, they will pay costs if the costs seem nominal at first.
If someone can’t write well and has an upcoming deadline for an essay or needs to write an important document, they might use a one-time fee. If someone uses it for day-to-day decision making, they will certainly do a subscription if it seems cheap at first. And once you have those people expecting to pay, the price keeps increasing every couple of years.
Example a - Facebook. Rolled out as an entirely free service and enrolled billions of members. Now enough people pay for extras in things as stupid as Candy Crush that they can tack millions onto their revenue for nearly no benefit provider. If enough people pay $1 for something stupid, it pays off
Example b - Prime streaming services. It was a free tack-on to people with Prime memberships. So people bought Fire sticks and use it as a home base. And now they are so cemented into the system and don’t know how to use an alternative so they will pay for 1) extra apps that will show the program they want, which will pay Amazon a fee and 2) pay extra for no ads now that they are addicted
I also have no doubt that advertising can be baked into AI. Just wait until it writes a pathology report and recommends a certain brand name medication, or gives someone a to-do list for hosting a Christmas party with specific brands and desired trends listed.
As for video generation, I think you have a small enough pool that does this anyway, and this will increasingly be consolidated by price towards a handful of powerful individuals who want to sway public influence.
You and I may not pay for it, but have no doubt you can get similar adoption to streaming services and you can get people to pay for anything if it feels cheap enough to be nothing money. I’ve become a cynical old fart and I’ve watched this song and dance too many times.
I think the price point is where your theory falls apart. The AI companies are spending far more than Netflix and they will need to charge more than the average family can afford in order to turn a profit. (Especially in the US where people are struggling to pay for gas and groceries right now.)
Take a look at the new token-based business services. People are blowing through their monthly allotment in a couple of days doing the same things they had been doing previously. Businesses won’t be able to afford increasing their AI spending by 10x to keep up.
I think it will have to consolidate down to a few players - I certainly don’t expect that there is a market for 100 different systems. And I think the personal market is a supplement to the executive suite market, so getting a few people addicted for cheap both brings in a small revenue stream and dumbs a population down enough that they have no choice but to function without it (imagine telling someone in 2001 that they would one day pay $1700 for a cell phone - now it’s just considered the cost of living)
The businesses who overdid it on tokens have a choice to make: hire back workers or switch to a different AI company. Maybe one day they will look at humans (although salary may have to increase due to negative reputation), but I think there are a good number that will just shop around to another AI company for a better deal.
There’s just no way to pay for the cost of these services, though.
When someone constructs a 100 MW data center (now considered a smaller one for new construction), that’s about $2 billion in total costs to outfit the whole operation. And then once it’s on, we’re talking something like $10-20 million/month in electricity alone, and a few million in other costs. How many $20 subscriptions do you need to sell just to break even with your operating expenses? How many $100/month subscriptions do you need to sell to make a dent on your interest payments on the construction? Will there be a market for $1000/month subscriptions from millions of customers? If not, how’s this all going to be paid for?
They certainly cant make their money back writing personal emails and doing kids homework, but I dont think thats where they are aiming in the long run. We may end up with big business and military paying big money for the real frontier models and everyone else using lightweight local models on their own hardware, cheap models integrated into existing applications, and watered down frontier models on subscription.
You may be right (I hope you are). Do you think government funding would offset this?
I don’t think government funding can actually offset the crash in consumer and business demand being insufficient to cover the cost of the most expensive models on the most expensive GPUs. But if you look through my comment history I’ve made the comparison to supersonic flight, because I genuinely believe there’s a possibility that governments fund the expensive branch of this technology for their own military or surveillance or law enforcement purposes without the benefits necessarily actually spilling out into normal commercial applications.
We’ve hit the point where training a model (both pre training and post training) isn’t the expensive part, and the expensive part is actual inference, which makes it hard to scale the most expensive models to where it’s useful for a lot of people. So it might be that the companies and governments that can afford to operate an expensive model might be the only ones to do it. And they’ll be able to, without necessarily the public being able to have access to the same tech.
They would raid pensions before asking congress
What happens when its chatgpt, Claude and Gemini $$$$ vs Deepseek $. The Chinese will absolutely flood the market with free AI so long as it burns the American companies.
With more headlines like this in the news …
Anthropic/OpenAI may be spending more than $1000 for every $100 you pay them
… it really seems like even if the market is winnowed down to the major players, there’s still not enough consumer demand at the prices they need to be profitable.
Also, if my cellular plan cost me $1700, I’d be shopping for a better deal.
Advertising is already baked into AI. There just need to switch it in fully. That’s coming.
People will pay $10 a month for chat GPT to write their emails.
They won’t even pay that in many cases but instead rely on the free offerings. Expect those to go away altogether, to be heavily hamstring, or to become absolutely riddled with advertisements.
If run locally you get none of those problems.
Right, but then we are back to the question of how these companies are supposed to make money. I suppose hardware manufacturers will be okay with this outcome but I’m not sure how these heavily inflated software companies will profit under a local first world. So we are back to the bubble deflating or bursting.
I agree with your conclusion. I can’t see where the income is coming from to fund all the datacenters being built.
The bubble is based on creating AGI, but people already don’t want to pay for sub AGI intelligence that is already beyond most people’s needs.
On the code development front it’s even worse as now you have an unpredictable cost based on token consumption rather than the predictable cost of a salary and have no leverage to negotiate the cost.
Quite true. Switching providers is an apples to oranges comparison, so it’s not like you can just switch from Claude to Gemini and expect everything to work perfectly.
I’m excited to see the PE Ratios after these IPOs. We’re already well into dot.com absurdity, but throw in some multi-trillion market caps with negative earnings?
Really? Has anyone mentioned that to Tesla? Or to whoever is in the government that keeps giving him billion dollar contracts?
Oh you mean the Tesla with a PE Ratio of 382 (20 times the industry average), a PEG of 16, P/B to 18, massive insider stock dumps, and is still down YTD? Tesla still has power because of their stranglehold on charging. AI companies have insane costs and people souring on them as they either lose their jobs or see their companies’ costs go up as a result, all without solid ROIs and strong buyers remorse.
And Trump/Hegseth keep giving them contracts because they’re idiots
still down YTD
TSLA was at $297 this time last year and it’s trading at $403 today. I’ll spot you that it’s off the $485 peak. But it’s been on an absolute tear for the last six years.
And Trump/Hegseth keep giving them contracts because they’re idiots
Government contracts are the lifeblood of every big corporation. Half the reason why business execs at Fortune 100 firms are so slavishly loyal to whomever happens to be running the country at any given moment.
I don’t think Trump/Hegseth are morons for giving Musk access to the unlimited money pump, given how much of Musk’s money keeps overflowing into their own pockets. I think Biden and Obama were rubs for pumping up Silicon Valley stock over the last 18 years, fully ignoring how often they were getting in bed with outspoken fascists.
More broadly speaking, I think Hegseth has set the Pentagon up for failure over the next decade in the same way Rumsfeld did under Bush. But I also don’t consider that a bad thing, broadly speaking. Just a shame that we’re going to piss away trillions in blood, sweat, and tears at home so the US can finally get its nose busted properly in the Middle East.
And Trump/Hegseth keep giving them contracts because they’re idiots
Ironically, Trump said the same thing about Biden.
Agree with the sentiment. It’s not like we are not in a bubble, but the old fashioned business and growth of the 20th century is gone. With technology things scale incredibly so winning the market share can put you on top for decades, and not just with finances. At this point the power and control of information means everything
Ah yes, the stable and sane business environment of the 20th century, which brought us totally sane and obviously correct investments like Xerox.
Xerox was a household name !
The only thing more guaranteed than a Tech Sector crash is the multi-trillion dollar government bailout.
Might as well walk in the door with a $1.5T valuation, because that’s what the Feds are going to use to justify cutting you a check in another five years.
Is the real hope that AI will be a subscription service? Those subscription billionaires ain’t got a clue. Let’s change “At some point you’ve got to make money” to “At some point you’ve got to add value”.
Based on what? Spotify took what, 16 years to turn a profit?
Money stopped working like you expect it to a long time ago.
Come on now, Nvidia is making money. Nvidia then just gives it back to them to make more money is all.
Yeah, Nvidia invests in Open AI, Open AI uses that money to buy Nvidia chips. More people invest in Nvidia because it’s revenue keeps growing.
Charles Ponzi would be proud.
Alan and Beth, two economists are walking through some woods. They come across s pile of shit and Alan says to Beth “if you eat that shot I’ll give you a tenner”
“Okay,” says Beth. She eats the shit and Alan hands her a £10 note
They walk on for a bit and find another pile of shit. Beth says to Alan “now I’ll give you a tenner of you eat that pile of shit”
“Okay,” says Alan. He eats the shit and Beth hands the £10 note back to him
They walk on a bit further and Beth says “you know, it seems to me like we both just ate a pile of shit and nobody gained anything”
“Ah, that’s where you’re wrong,” says Alan “we’ve increased the GDP by £20”
Isn’t this just the broken window fallacy but with a poop fetish?
Not quite. The point of the broken window fallacy is that had the money not been spent on the window, it would have been spent on something else. So breaking a window does generate revenue, but not necessarily more than not breaking a window
Whatever happens, the shopkeeper is in a worse position than he otherwise was
Here the point is that they’re passing the money back and forth. That’s like the AI companies right now - passing money round in a circle and that being presented as if it were the same thing as the industry as a whole making money
NVidia gives money to OpenAI, who give it to another company to build datacentres, who give it back to NVidia and NVidia’s Number Goes Up
But wouldn’t that be velocity of money, not GDP? GDP is a measure of final goods and services, not money. Unless you have a poop fetish and consider the eating of poop to be a final service, then this wouldn’t impact GDP.
Here the point is that they’re passing the money back and forth. That’s like the AI companies right now - passing money round in a circle and that being presented as if it were the same thing as the industry as a whole making money
The stock market value isn’t the same thing as GDP though. The GDP measurement is never really a great measurement of economic well being because of inequality, but even then, we really don’t know what it is in the US because Trump fires anyone that gives out a bad report.
The people handing money back and forth quickly was actually what happened when the pandemic ended, the increase in the velocity of money resulted in inflation but people blamed it on Biden. Now Trump is President, the economists have been fired and fraud has been decriminalized.
Wasn’t his full name Charles Ponzi-Scheme?
wow these guys seem to hate it all of a sudden, what changed…
They road the wave up, now they are wanting to exit before it crashes
Exactly that. If they do an IPO, they either will continue to waste the new shareholders money like they did with their own, or they actually have to make money, which would imply to raise prices at least ten- to twentyfold while hoping to keep all customers on board.
I don’t think either of these ways will go.
I’d like to know what they are promising in their IPO prospect…
Times running out? No, it’s out.
The market can remain irrational longer than you can remain solvent.
It doesn’t seem like you do anymore
I want to see everyone hop on the IPOs and ride them all the way down to the floor!
Fuck AI
I think investments, 401Ks and retirement funds will hop on SpaceX and AI and the mass will complain for a bit until the next news cycle.
If I wasn’t convinced the entire financial system was fraudulent I would be shorting the fuck out of all of them
“The market can remain irrational longer than you can remain solvent”
We’ll burn more and more coal to power air conditioners to deal with unbearable global warming until that’s no longer possible.
The financial markets are the same. No reasonable person believes any of this is sane or sustainable. But what happens when the music stops? Very few chairs.
Especially since Nvidia has reached the “too big to fail” level; guess who’s gonna pay again for the upcoming crisis?
Source? Not doubting you, just would like to know the criteria.
There isn’t a completely objective criteria, from Wikipedia:
Federal Reserve Chair Ben Bernanke also defined the term in 2010: “A too-big-to-fail firm is one whose size, complexity, interconnectedness, and critical functions are such that, should the firm go unexpectedly into liquidation, the rest of the financial system and the economy would face severe adverse consequences.”
NVIDIA currently makes up nearly 7.5% of the S&P 500 (not to mention that they have invested a lot in the other companies and those companies have invested in NVIDIA). If they were to suddenly be liquidated I’d say that would cause ‘severe adverse consequences’ to the economy.
And it would likely cause a ripple to other tech companies, which which make up roughly 1/3 of the entire US stock market.
Thank you indeed. On my side I only remembered the news back in November that nvidia weighted 16% of the US GDP. Now of course comparing valuation and GDP is not really meaningful but the idea stuck.
Thanks, especially for the figures, that helped a lot.
‘At some point you’ve got to make money’
Hey hey hey, now, that’s loser talk. Here, c’mon, let’s snort a ton of coke.
That sounds like something The Wolf of Wall Street would say
A lot of people working in finance were not alive when that movie was taking place and like the movie Wall Street the ones who watched it learned all the wrong lessons.
Don’t you worry about Wall Street, let ME worry about blank!
Blank? BLANK??! You’re not looking at the big picture!
