The AI Boom May Be Sowing Seeds of Another Financial Crisis
bloomberg.comI lost my job during the dot com bust (it preceded the GFC) and since that shook me, I was very skeptical of properties making people rich overnight but they literally were!
"There's no free lunch!", I told myself, but it was becoming increasingly hard to ignore my barber, the cashier at the grocery store, my dentist, my landscaper - all who had bought highly leveraged properties and had actually made profit. I'm not ashamed to say I started to look into it myself - I was even ready to purchase some leveraged properties myself but kept getting outbid by even crazy offers.
Then 2008 happened. It happened suddenly. First it was on the news. Then quietly, people began to stay at home. Then move out.
It was both unsettling but also reassuring. My home value halved overnight (it's 800% up now so I'm doing OK) and mu neighborhood went mostly empty (bankruptcies, foreclosures).
Now again in 2026, my barber, the cashier at the grocery store, my dentist, my landscaper - all cannot stop talking about Grok Bot, about "Chat" (they usually mean ChatGPT or Gemini).
Maybe this time I can get in and get out in time, get real rich!
Maybe this time it's different!
Are they talking about using it as a consumer? Or somehow getting rich off it? I think those two indicators are very different.
> Are they talking about using it as a consumer?
Consumer. Their usecases seem pretty valid - the landscaper is using it to drum up more business, has agents responding to incoming emails and texts, just like my dentist and barber.
The cashier uses GrokBot to transcribe notes to her manager. Her company has a very bad inventory app (I don't want to name names but it's a very large nationwide chain) so she literally exports inventory into an excel and feeds it into GrokBot as a way to run queries.
I can see nothing wrong in them just like I could see no wrong in buying properties and flipping them for a profit because there was so much demand at that time.
If you weren't buying houses 2003-2009, you have no idea how crazy it was. People would buy homes unseen in Olathe, KS.
It's likely just me but it just feels eerily familiar. I hope we are all wrong this time around.
uh, there's a huge difference between using an AI agent(s) to make yourself more efficient in your line of work, v.s. buying leveraged properties valued in the hundreds of thousands of dollars.
Let's get the real indicator here: how full are the parking lots in the Bay Area these days?
Both have zero ROI in a rational market.
Why would a rational market assign zero ROI to a landscaper managing their communications more efficiently?
because, in a rational market, if a competitive edge is gained from a 3rd party service (GrokBot) that won't refuse to sell to the landscaper's competition, then the landscaper's competition will also use GrokBot the same way where that competitive edge is no longer there.
In other words, all the landscapers are equally efficient and unless they collude on pricing, the then pass on savings to their customers.
Which means any landscaper who doesn't use "GrokBot" will lose customers to those who do and have lower prices. For every landscaper, other than the first one to adopt "GrokBot" and lower their prices, there is ROI.
correct. As long as one landscaper has an edge, they will likely have a ROI for that window of time. However, when you calculate ROI, you don't do so over weeks or months, but rather years unless the specific job itself is very short lived.
For back office operations, one calculates ROI over 5 years or so as upfront investments are rather large (less for turnkey SaaS but that's a whole another discussion and not what's happening here. Here the owner is building their own solution).
Your theory is sound. What you're missing is in a rational market, 3rd party edge only provides an initial boost until the rest catch up. That's why companies invest in in-house talent, so they own and control it.
I think one of the things you're missing here is that landscaping (and which grocery store you go to, and which barber you go to) is not really a "rational" market.
People stick with service providers for a long time. The lead that the landscaper in your example is building now actually compounds over time, regardless of whether competitors adopt grok bot in the future or not?
In your 2008 example, making a profit was contingent on cashing out (selling your property). Whereas the landscaper is expanding their book of business and increasing their future cash flows.
Unless you think there will suddenly be some sort of rug-pull in his ability to use AI tools to manage his comms, or a sudden change in long-standing consumer behavior around landscaping services... I don't understand how you can in good faith think these economic scenarios are similarly risky? What debt/leverage is the landscaper taking on by using grokbot?
Great questions!
> The lead that the landscaper in your example is building now actually compounds over time, regardless of whether competitors adopt grok bot in the future or not?
You are conceptually right except that you are conflating cost of service with a business relationship.
If my business relationship offers me the same cost of service for the same service as their lower priced competition does, it's then not even a thing for me to worry about. They are completely financially fungible (plus the value of the business relationship).
Assume I work with landscaper A while landscaper B is using grokbot to be more efficient.
If my business relationship with landscaper A is strong, I will happen to mention to them what landscaper B is doing and if landscaper A values our relationship, they will try their best to do so as well or comp my service accordingly.
There are other factors like churn, overhead, regulation, licensing etc that I am happy to discuss in more detail.
Look at the graveyard of ChatGPT wrappers. Same concept.
This is why a grandparent observed that amortized ROI that depends on a 3rd party service becomes 0 - in capitalism, an arbitrage opportunity is competed away.
This is why companies build their own in-house talent for their core competencies - so that their ROI on their core competencies doesn't become 0.
This is also why sama and Dario are trying to regulate OSS models - because those are a direct attack to their core competency and if left unchecked would cause headwinds to their ROI,
In 2008, your landscaper went bankrupt because they took out a sub prime mortgage.
How is your landscaper going to go bankrupt in 2027 because they're using grokbot?
I forgot another source of ROI. As each landscaper reduces their prices their addressable market (people able to afford landscaping) expands. By saving time using AI, they are able to serve more of that market. So the actual ROI will still be positive.
And keep in mind this is one market with relatively undifferentiated players (landscaping services) that we are talking about. Markets where operators have more pricing power will be even better off.
If we follow your argument to its logical conclusion, any technical or capital investment by a landscaper is pointless. They may as well use manual lawnmowers, rakes instead of leafblowers, and go from job to job on bicycles instead of pickup trucks.
> By saving time using AI, they are able to serve more of that market. So the actual ROI will still be positive
Not really. Most services and labor isn't elastic. For one, you cannot service more than a dozen people a day if each service takes 30m. To scale, you will then need to hire more people who will want to work on a salary and will demand healthcare and days off.
But if AI saves you 30 minutes a day you can serve an additional client.
if the additional client is literally next door and you don't need a break, refuel or your batteries didn't run out of charge, yes
Your property being up 800% is not you doing ok.
It is a sign of a deeply broken system.
> It is a sign of a deeply broken system.
Yes it is. Pulling a simple permit takes months and thousands of dollars. Building a new house is practically out of the question.
The SF Bay is deeply supply constrained but just because tech billionaires here can buy up a dozen homes while offering millions over asking, the city gets their taxes and so they are not incentivized to improve things.
Isn't there a story about Rockefeller and a shoe-shine boy with stock tips? Because that seems apt.
I believe that was Joseph Kennedy, father of JFK and RFK. He realized that he didn't understand what was going on in the stock market at all, and so he got out, a few weeks before the 1929 crash.
Off topic: I first wrote "2029" instead of "1929" because habit. Wow, the start of the Great Depression was 97 years ago? I find that surprising.
Joseph Kennedy did not get out of the stock market, he turned around and sold short. In 1929 he was worth $4M. Five years later he was up to $180M.
Well what did you and them do after "getting rich"?
That helps us decide whether there is some purpose and meaning to the entire story about the system.
> Well what did you and them do after "getting rich"?
I never got rich as I never ended up buying a property for investment. I just kept on getting outbid by others. I kept getting outbid because I didn't know how I would pay $1M for a home that I would struggle to rent out for more than $800/mo.
The others reinvested their profits into the next purchase until the music stopped. Everyone but my dentist got wiped out.
My dentist just made a payment everytime he got a late notice which was like once every 4-6 months as the note changed hands, often getting lost in the chaos. He refinanced all the properties around 2012-2013 and invested the proceedings all into one multi storey clinic which is doing very well.
a lesson there is definitely relevant for Y Combinator startups (and I assume YC teaches): know your actual obligations, not what the people you're obligated to wish they were. Avoid spending large amounts of money unless you actually have to.
> Maybe this time I can get in and get out in time, get real rich!
Well, since you seemed to know what you needed to do back then (but didn't take action), what do you think you need to do now to "get rich"?
> Well, since you seemed to know what you needed to do back then (but didn't take action), what do you think you need to do now to "get rich"?
You give me too much credit. I didn't know how to "get rich" back then and neither do I, now.
I didn't take action because I didn't know. I kept getting outbid because I didn't know how I would pay $1M for a home that I would struggle to rent out for more than $800/mo.
ah okay, my mistake!
My answer, as always, is: Organize as workers against our capitalist exploiters, and overthrow this system of capitalism in favor of a system which rewards us workers for the wealth we create and not our capitalist exploiters who’s only contribution is rent seeking on the wealth we create.
Capitalist countries are the ones with the highest standards of living, and formerly communist countries like China gained immense wealth with pro-market reforms, so perhaps the reason people on HN may not support a revolution is that they're actually more informed on this inconvenient reality than you assume.
The market reforms got China’s GDP from 1000 CNY to 7000. The communist reforms got China’s literacy rate from 20% to 80%.
I know you capitalist love money more then anything, but the rest of use like to live healthy lives and enjoy things like good books and spending time with our friends and family.
> but the rest of use like to live healthy lives and enjoy things like good books and spending time with our friends and family
If you're in the U.S., what's stopping you from doing that? Some of my coworkers in tech retired to the country, raise their own chickens, goats, grow their own food and are all on solar. They seem to be loving their life.
What is stopping most of us? I suspect it is not winning the start-up/ipo lottery.
> overthrow this system of capitalism in favor of a system which rewards us workers
This time it will be different.
They call them revolutions because they come round full circle.
...and yet you're hanging out on a messageboard that's operated by a VC firm for commercial reasons.
What precisely are you accusing them of? As a gesture of good-faith, I'll put my subtext-cards down on the metaphorical table first: I believe the quippy-looking retort is actually a dud once we open it up and look inside.
For example, is it one of these?
1. "Being a regular here necessarily means you cannot actually have that opinion, so you are lying to us."
2. "You may believe that, but posting it here is not logical because nobody here will be receptive to it."
3. "You may believe that, but posting it here is an immoral offense against the management."
Everything on this board is controlled for the purpose of making billionaires richer. None of your three things are true. Here's 4: "the fact that you are posting about the revolution on Make Billionaires Richer Inc and not on Anarchist Graffiti Dot Onion is a sign that the billionaires have thoroughly won and the revolution will not happen"
The revolution won‘t happen on Hacker News, that is for sure. However, Hacker News is one of the biggest social platform of tech workers, and GGP might not like it, but when so many workers in a single industry come together, there will be solidarity, particularly against our bosses. And among workers in any industry there is always a radical (usually me) talking about the revolution.
I'm not accusing anyone of anything. I'm pointing out an irony that amuses me.
From the anger in the replies perhaps it touched a nerve, eh?
The only anger I am sensing in the replies, and the only anger I feel towards your post, is about the cliché debate tactic “posted on an iPhone ... gotcha”.
NFTs. I was listening to this one podcast The Million Dollar podcast where they tried to turn 10k into a million dollars. They started a bit late and lost most of it when NFTs and NBA trading cards lost their value pretty quickly. They had some really smart-sounding theories about great stock picks to make boat loads of money too!
Ponzi schemes are extremely profitable if timed correctly. But statistically, it's impossible for the majority of people to have correct timing. You have to be twice as likely to lose $200k as to make $400k because the whole system averages out to zero.
If you have the right heuristic for when to get out of the scheme you can become an overnight millionaire. I don't have it. Statistically, few people can.
> Now again in 2026, my barber, the cashier at the grocery store, my dentist, my landscaper - all cannot stop talking about Grok Bot, about "Chat" (they usually mean ChatGPT or Gemini).
I mean, they should be. It's a transformative technology that is doing something that was science-fiction five years ago, and it going to make a few people trillionaires, and it's not entirely clear where it will leave the rest of us.
That doesn't mean that you should, like, be rushing to be a bagholder.
> That doesn't mean that you should, like, be rushing to be a bagholder.
I think the challenge is, in 2026, no one has a clue what the bagholding will look like
Being a normal, working-class citizen.
It’s a bubble. It’s way bigger than the dotcom bubble. It will burst, and it will take trillions and trillions of dollars of value with it.
With any luck, all the billionaires in the world will jump out of their windows, just like they did in the ‘29 stock market crash.
October of ‘29. Just three more years to wait.
I came across https://www.youtube.com/watch?v=q2nA2szz8dY a few days ago from during the financial crisis, and the things he listed as the four causes all apply here too just replace Property Developers with businesses adopting AI as part of the hype cycle. The debt and capital markets have their part to play this time too, just not the primary problem. But the ingredients are all still there for AI.
It's absurd we're in this position to begin with. Bloomberg taking years to find out what technical people saw months if not years ago.
News media isn't a truth-finding machine, it plays for engagement. If Bloomberg is publishing this article now it means Bloomberg thinks publishing this now is good for Bloomberg's bottom line, and nothing else.
So it is of course AI being the one creating the financial rather than crypto.
Makes the binance crypto incident look like nothing (and everyone was made whole).
The hidden $3.2T AI debt on the other hand...
May? I'm pretty sure there's a crisis coming one way or the other, but the crash happens only when they fail to make God In A Box.
The goal of God In A Box wasn't a good goal anyways.
Most people can't agree on what God In A Box (AGI, the singularity, whatever) even is. We don't have a widely-accepted definition for it.
The SMART criteria for creating goals has been around since 1980 [0]. Goals need to be Specific, Measurable, Achievable, Relevant, and Time-Bound. This is stuff that they teach most people in bachelor degree-level college course. Not even. Maybe in high school or associates degree-level courses.
If God In A Box was the goal that SV sold all of these investors on to the tune of over a trillion dollars of investment, the best thing we could do with everyone involved is have a massive settlement where all of their money just goes to pay off a fraction of the national debt and they're banished to live in the Alaskan wilderness, never allowed near more than $50 at a time or long-distance telecommunications devices again for as long as they live.
It could be worse. There's a chance they succeed.
I'm pretty sure that the worst things will not come from AI/tech, but from other things – education freefall, population crisis, pension system collapses in Europe etc. If at all, AI and the current situation in tech is one of last false hopes to ignore all this.
AI is going to be a driving force behind it if it begins to work. What do you think the results of most humans suddenly becoming economically irrelevant will be?
Keep in mind that today, if essential workers go on strike, those in power have to listen. What happens if people on UBI refuse to... something? Especially in a world of increasingly automated policing and surveillance?
> What do you think the results of most humans suddenly becoming economically irrelevant will be?
Interesting times.
When people on UBI refuse to not riot - see France.
Yes, why do you think there's a massive investment into police tech today? The bets on what the future will look like are getting placed.
Uncensored local AI + kitchen chemistry + 1 million discontents.
I think the humans still have quite a chance to overthrow a police state tyranny.
If it was that simple someone would have already done it. Even in places like the USA, attempted political assassinations are very rare.
But what you say also gives us a path back to economic relevance: you can make money by selling policing equipment to increasingly paranoid states. They'll never be satisfied with what they already have. It doesn't even have to be a useful product, it just has to satisfy their paranoia while also fueling it.
Sounds like a great future where everyone prospers.
If Ai works there is no reason there has to be a pension system collapse. I think the problem will be complexity. Ai will allow such complex systems that fewer and fewer people will be able to maintain the systems because the complexity will be so high. We'll need to plan for what it looks like that people with IQs less than 145 won't be productive in any meaningful sense of the word.
High intelligence is overrated. Medieval knights didn't dominate the battlefield because they had the potential to become the best fighters. They dominated because they had access to weapons, armor, and horses, and because their social status and wealth let them focus on combat training.
What is education free fall?
China is ready to step in with millions of PhD-level graduates.
Well if everyone starts just asking the LLM to write their essay and ask it for the answers on their math homework...are they really learning anything?
Anyone who truly buys into that I pity. But it's sure a useful line to help sell the magic.
Well, either it's going to fail or the trillions going into finding breakthroughs are going to find breakthroughs. How much gasoline do you want to give arsonists, even if at the moment they're only pouring it onto concrete?
I would like to take away the gasoline before they figure out that wood burns better.
1. Hedge funds buy 2. Hedge funds exit 3. Pension funds buy 4. Early investors exit
"May be"? I think everyone sees this one coming.
I agree that there's much wild speculation, but some of the speculation is pretty mild in comparison. Micron trades at a forward P/E multiple of approximately 4.3 x to 5.0 x based on consensus forward earnings estimates. Those aren't bubble numbers. Those aren't even utility numbers anymore.
I think everyone is worried and, at least in this case, being cautious.
Countrywide traded at like 6x earnings in 2005. The market not putting a high multiple on peak earnings for a cyclical is better than the alternative but it doesn’t imply there’s no bubble.
And because everybody sees the risk, it is already priced in.
The stock market is unpredictable precisely because everybody is trying to guess what is going to happen next. Dollar cost averaging takes the guessing away.
Nothing is ever really priced in when it comes to layered structured debt and credit default swaps. Not even the default swaps.
Go on and make millions from mispriced assets then.
Indeed. Not sure how DCA comes into this though.
The bubble can't pop until most investors believe it's not a bubble, it's the new normal. As long as they believe it's a bubble they'll have money on the sidelines to throw in whenever it goes down.
Or at least it can't pop for purely financial reasons. It might still pop if, say, Anthropic and OpenAI both announce they've hit the scaling wall, or Iran somehow bombs TSMC, or a court rules they have to pay a royalty to every book author every time they answer a prompt.
> Not sure how DCA comes into this though.
Developing the habit of investing at frequent regular intervals eliminates the desire to time the market. And that's what threads like these are about, for most people.
Not until one of the major AI companies says they’ve hit a limit for intelligence for models.
I think the system is more fragile than that. Infra providers even now have to raise money with junk rates meaning that are entering a debt spiral. If that compute doesn't come online because the infra providers are bankrupt Anthropic/Open ai can't hit their growth projections and Nvidia sales will drop as well.
And I am not even speculating if the growth/revenue projections are realistic in the first place, which, if not, would be another trigger
The markets appear to be hitting their limits first, and it’s the money that’s keeping this show on the road. Why do you think Anthropic is racing like mad towards an IPO?
Goverment regulation... when goverment has regulated wall street and banking industry no issues... as soon as "we will let them regulate, no mam, door plugs will not start failing on airplanes..." things go south.
Looking at GPU prices today, this is an understatement.
My finances are in crisis.
You know whos winning ... the bitcoin miners retiring unprofitable cards, if they still have them.
Good job guys.
> You know whos winning ... the bitcoin miners retiring unprofitable cards, if they still have them.
Bitcoin mining hasn't used GPUs in ages. They've been all-in on ASICs since 2013 or so; even if the GPUs were free, they wouldn't be worth running.
Bitcoin proper, yes. But not Ethereum; that bubble didn't pop until 2022-ish. The hardware constraints of ASICs made them useless around 2020, causing yet another surge of GPU usage.
Yeah, but how logical is that when you don't even have the electrical grid capacity to turn on all the GPUs that have already been bought?
I would think if they were mining with them before, they had the capacity at that time.
After bitcoin halfed enough, the electricity cost more for the miners.
Now with the AI boom, people are paying premium prices for low end cards, because they run AI tasks exponentially faster than CPUs do.
So the electrical load that was once centralized at the mining site, gets distributed.
Being able to schedule your AI agents to work during off peak electrical prices, and sleep during peak pricing, seems to be a stablizing factor for the grid.
Im talking about the used video card market.
Not the billion dollar data center market.
Prices are up 35% in the last 3 months alone, and I just paid way too much for a RTX 5060 Ti 16GB for the ram.
@angry_impala .. thats their business. i needed a card. :|
Adapt to the worlds problems.
Don't adopt them.
I get it, I'm just saying when people point to GPU prices as some sort of proof this is all logical, it's actually the place where the logic begin to unravel the most blatantly. They have more GPUs then they will be able to plug in for at least a few years.
This is why data centers are starting to burn fossil fuels on site behind the meter for power. Including using diesel generators.
The newer data centers aren’t relying on the grid. They’re more likely to be selling excess power than buying it.
I remember there being a run on GPUs during Covid for whatever blockchain nonsense was going in then
Sowing the seeds or failing to paper over? Pretending this is new is baffling.