-
After the bubble pops how much would our lives be impacted?
-
Would AI vanish or still be there?
-
How exactly do you think the bubble will pop? Will AI companies simply run out of money? Or will it be because of the environmental effects?
-
When do you think the “pop” will take place?
-
After the bubble pops, in future there will be companies/people who will try the AI thing again? What will that be like?
I think part of the process is going to be the enshittification of AI services. (Temporary, until the technology for sustainable infrastructure catches up.) Tokens will cost a premium, and a lot of businesses that depend on AI will turn to China’s cheaper offerings.
Hobbyists will turn more to home-grown AI systems.
In recent news, Chinese AI is catching up to Anthropic. It’ll be interesting if the Trump administration tries to block the Chinese AI offerings in order to bolster US ones. Trump and the AI tech bros are familiar bedfellows after all.
You’ll be able to buy LLM hardware (video cards and RAM) for cents on the dollar as companies fail. The NASDAQ will drop a bit, the number of LLMs will reduce
Better than the GFC where all we got was Enron corporate branding and expensive houses
Comparable to dot com where loads of nerds bought $5 servers
After the bubble pops how much would our lives be impacted?
A bunch of pensions will lose a lot of money. The billionaires will know when it’s the right time to short it, and so they’ll get wealthier.
Would AI vanish or still be there?
Higher end models like Mythos will still exist but be too expensive for most people to use. Lower end models will still exist. People will start developing actual tools focused on specific tasks using LLM algorithms instead of AGI. Companies will be able to buy LLM appliances they can put into their server rooms that will help with data analysis which doesn’t require sharing data which will be more useful for use with information they want to keep private.
Eventually people will have PCs, then laptops, then phones with chips optimized for LLM algorithms which don’t require sharing private information with a data center. It won’t be AGI, but it will be able to automate things on the computer and help with stuff like helping with grammar. People will be able to say something “computer, bring up the financial report from last week” and it’ll be able to do that. But it’s not going to be an AGI thing, just be able to find the correct program, files, etc. based on some simple cues.
How exactly do you think the bubble will pop? Will AI companies simply run out of money? Or will it be because of the environmental effects?
Similar to how the dot com bubble popped. Some chatter from billionaires about getting out results in them all shorting it at the same time, the share values collapse, but they make a bunch of money.
Will AI companies simply run out of money?
Yes, when they’ve gotten the last investment dollar, the billionaires will know it’s peaked which will result in a big short.
Or will it be because of the environmental effects?
They don’t care about environmental effects LOL.
When do you think the “pop” will take place?
You need to be invited to whatever is the present day equivalent of Epstein Island to know that.
After the bubble pops, in future there will be companies/people who will try the AI thing again? What will that be like?
Yup. We didn’t stop using the internet after the dot com bubble burst. It’s just the financials of it wasn’t insane anymore. LLMs can’t do everything, but it’s an algorithm that’s useful in some scenarios. It will still be useful in those scenarios, but people will stop being weirdos thinking that it’ll be able to do everything. It’s an algorithm, there’s a cost for the hardware and power to run it, in many cases it won’t work at all, in many cases it’ll work but be cost prohibitive, and then in a few cases it will be actually useful and cost effective. We’ll move to using it for only the latter case.
Over time, it will get better and be used for more things, some of the things they’re promising now may become a reality in a few decades. Kinda like the ridiculous idea to sell dog food on the internet during the dot com bubble. It was ridiculous then, but eventually the logistics caught up and more than a decade after the dot com bubble you could indeed buy dog food on the internet. I imagine it will be like that with AI. Eventually some (though not all) of the things they say you’ll be able to do in six months may actually happen… 15 years from now. But it will probably be enshittified and suck more than you imagined it would be.
The only piece I think you missed is that nation states and surveillance corps will buy up foreclosed data centers for pennies on the dollar to use them to power a surveillance/police state.
You’re assuming they don’t already have a copy. All of the AI companies have government contracts after all.
Our 401k will go down a bit, a bit of money will be transferred from retail investors to the big players, the rich will get richer and the poor will get poorer. You won’t see so many ads for dumb AI services that literally no one wants. The big players will be fine, maybe, maybe, a few data centers shut down, but most of them don’t.
Anybody thinking AI goes away is just wishfully thinking, and/or ignorant.
It will if it is too expensive to generate a positive return on investment. That’s not to say that there aren’t (very) specific use cases where running an LLM makes sense, but the general use case of attempting to replace human workers with AI will likely be recognized for the pipe dream that it is. Companies are already realizing this with the AI bills coming due now, and this attitude will compound once they start having major failures and no one is still around who can fix the problems.
Are you saying that you think there is a possibility that AI technology, especially LLMs, would disappear from usage? Even as search tools? Even used locally?
Like I said above, I think there will be some very specific cases where running an LLM makes sense. Think along the lines of querying lakes of scientific data. This won’t be something of value to the average Joe or the average company. My experience with using LLMs for general search has been very hit or miss. I always end up using a general search engine, even for internal data sources.
Maybe that’s just you but millions of people are interfacing with LLMs every day as a replacement to a traditional search engine. I don’t see that changing. Either way, sounds like you are agreeing that AI will not go away entirely. That would be a silly thing to say.
It’ll also go dormant as LLMs are only one AI tech. Other AI tech will be developing while supporting infrastructure grows until we have another boom. World models for example and JEPA. That’s why Meta is pushing AI glasses. They need training data.
Those millions of people using AI for general search are currently getting it for free, which is going to end when the bubble bursts and the companies providing it go under, if not sooner when said companies try to make a profit. This is exactly what’s happening to corporate AI customers now. They are realizing there isn’t a good ROI now that OpenAI and other AI providers are being forced to charge proper costs for queries and tokens, rather than the teaser or “adoption” rates they charged before.
While I agree that AI is here to stay, he did say the bubble popping. I could believe that there could be a reduced level of investment for a while. In the past, we did have periods where we thought that various AI tasks would be easier to solve than they were, and investment fell back as we discovered that there were more hard problems to solve before we could accomplish a particular feat. Didn’t go away, but did see a decrease in work on it for a while.
https://en.wikipedia.org/wiki/AI_winter
In the history of artificial intelligence (AI), an AI winter is a period of reduced funding and interest in AI research.[1] The field has experienced several hype cycles, followed by disappointment and criticism, followed by funding cuts, followed by renewed interest years or even decades later.
The term first appeared in 1984 as the topic of a public debate at the annual meeting of AAAI (then called the “American Association of Artificial Intelligence”).[2] Roger Schank and Marvin Minsky—two leading AI researchers who experienced the “winter” of the 1970s—warned the business community that enthusiasm for AI had spiraled out of control in the 1980s and that disappointment would certainly follow. They described a chain reaction, similar to a “nuclear winter”, that would begin with pessimism in the AI community, followed by pessimism in the press, followed by a severe cutback in funding, followed by the end of serious research.[2] Three years later the billion-dollar AI industry began to collapse.
There were two major “winters” approximately 1974–1980 and 1987–2000,[3] and several smaller episodes, including the following:
-
1966: failure of machine translation
-
1969: criticism of perceptrons (early, single-layer artificial neural networks)
-
1971–75: DARPA’s frustration with the Speech Understanding Research program at Carnegie Mellon University
-
1973: large decrease in AI research in the United Kingdom in response to the Lighthill report
-
1973–74: DARPA’s cutbacks to academic AI research in general
-
1987: collapse of the LISP machine market
-
1988: cancellation of new spending on AI by the Strategic Computing Initiative
-
1990s: many expert systems were abandoned
-
1990s: end of the Fifth Generation computer project’s original goals
Enthusiasm and optimism about AI has generally increased since its low point in the early 1990s. Beginning about 2012, interest in artificial intelligence (and especially the sub-field of machine learning) from the research and corporate communities led to a dramatic increase in funding and investment, leading to the current (as of 2026) AI boom.
Obviously, we did achieve a number of those — like, we have pretty solid machine translation of human language today. I remember, pre-Brexit, a senior EU translator for the UK talking about EU translation work. One thing he mentioned was that he did all of his first drafts via Google Translate and then just did manual cleanup by hand — and I’d call that a fairly prestigious translation position. But it took more time and research than we initially expected.
-
deleted by creator
The industry will end up with one or two winners and a lot of losers. Very hard to predict these things (smart money in 1997 was that Apple was about to file bankruptcy). One or two companies will have success and the rest will evaporate or be absorbed. Then the AI product itself will be priced up (and dumbed down) significantly.
AI will not go away, but will be far more limited and specialized. The day of free general access to powerful chatbots will be up.
I think it will be triggered by the massive overvaluation of many AI labs. Investors will sell. In the meanwhile, other sectors will still have to adjust to the new reality of token generators, so the wider economy is impacted.
LLMs will not disappear, just like real estate did not disappear after 2008,and the internet did not disappear after the dotcom bubble burst.
But I can imagine the end of hyperfocus on the “AGI race”.
I’m assuming one or two significant players will go bust, but probably not the biggest like Google, Microsoft, Amazon. They might take some serious hits, and they’ll deprioritize AI, but still exist. Maybe someone like Oracle or xAI goes under and gets parted out like an airliner getting scrapped. A lot of data center projects will be cancelled, others will shut down, and the local politicians that pushed for those data centers will see their political careers end as their districts are left with big empty buildings and infrastructure debts that will be hard to pay.
Wouldn’t such turbulent times be exactly when established companies are at risk as well? E.g., what electrification is doing to VW.
Although the risk for AI-first vs general big tech does depend on the root cause of the disruption.
Well yes, but for example, when someone activated voice interaction with their phone, Google or Apple decide what AI solution is used, and Google already decided to roll their own. When most companies go into AI, just expanding their existing relationship with Microsoft is what their business leaders like. Generally the established companies have preserved the customer engagement even when they resell newcomers.
An example of what you ponder would be like Sears, kmart, Amazon. The thing was that Amazon was greatly valued going direct. People loved the prices and shipping speed. Imagine if instead Amazon decided to partner with Sears and K mart and those companies largely handled the customer engagement with Amazon fulfillment on the backend. Amazon would have been much more exposed to those companies bringing it in house. This latter situation more closely resembles OpenAI and Anthropic right now, at least for the big businesses likely to pay enough after the price hikes.
like this. it won’t go away. it will oscillate and then stabilize
Internet was the “AI” of the dotcom bubble. It didn’t go anywhere after the bubble burst and neither will AI. What will go away are the companies with only hype but no product. Amazon and eBay came about during the dotcom bubble and now companies like OpenAI and Anthropic will likely be equivalent ones.
The bubble bursts when investors start losing faith that the company they have invested into will become profitable so they pull out, which makes other investors get scared so they pull out too and then the company goes under.
OpenAI and Anthropic will likely be the companies going under when the bubble bursts. Because unlike all the other big companies involved in AI (Nvidia, Microslop, Google, Meta etc.), they don’t have anything else to diversify with and they have never made a profit. When investors withdraw they are done for. The other companies will take a hit, but they will survive, because they have profits from other things.
The difference is that there were services that could leverage the Internet and be profitable. Every AI company is wildly unprofitable and burning unfathomable amounts of money. There’s no path for them to become profitable, unlike some companies in the dotcom era.
Exactly. The AI bubble popping will not “un-invent” AI technology. The technology exists and can be a useful tool for the right use cases, just like the internet still existed after the dotcom bubble popped.
LLMs will still exist and will still be trained and used. What won’t exist are AI only companies.
I personally would exclude OpenAI from my list of likely to endure. OpenAI partners are largely switching to Anthropic, their models and tools are generally less well regarded. The only time I had someone advocate for them it was due to some scenario where it was much cheaper.
Problem is that based on what I’ve read. Altman has way overextended his company. The financial obligations outpace their revenue and likely revenue way too much.
Anthropic has some risk since Microsoft, Google, and Apple insourcing would be a big problem for them.
Considering from where the bubble grows, my wisest guess as to when it will start showing cracks (and I’m not very wise, by the way) would be Republicans doing bad at next elections.
As of now it seems that any losses from AI industry would simply be tanked (Nvidia’s capitalization, for example, still ain’t a joke, they’ve got a lot of money to burn) or socialized. Little bits of regulations or opposition to socializing of losses may actually move the loss needle beyond pain point, where we can hope for panic to ensue.
If bubble pops - AI is here to stay anyway. Chinese have shown with their DeepSeek that you can run decent sized models locally, and the main gatekeeper of local AI is, you guessed it, Nvidia. Prices for server GPU’s are disproportionately high and one of the reasons is to not allow plebs like us to have that. But with further optimization of AI models they’ll become more and more accessible on hardware consumers already have.
My second guess of when though would be investors getting tired of AGI promises as it is AGI that has to magically turn money pit of AI into unprecedented profits.
The old message to people: only the first people get rich in a gold rush. The people who make the real money are the people selling shovels to the late investors.
Nvidia are selling shovels. They’ll do great as long as they prepare for the pivot after the people rushing give up.
GenAI is here to stay in one form or another, with long term impacts.
I wager the bubble will pop when OpenAI finally admits they have no path to making good on their purchasing commitments. They aren’t the whole bubble and Altman has made what should be obviously the worst financial moves, however the broader market will be more bearish on anything they vaguely think to be OpenAI like.
I suspect it’ll be with a year from now, based on what I’ve read.
The companies will never stop doing the GenAI things, but they may be less utterly obnoxious about it. I suspect Google and Anthropic have the highest chances of enduring a pop, anthropic for being well regarded in the field and Google because they get to make most phones go to them automatically whether requested or not.
I think GenAI will persist in obnoxious ways, but at least more bearable without folks desperately needing it to be adopted for the sake of their wealth.
The AI bubble will pop the same way the dotcom bubble popped. A lot of money lost and the hype will die down, but the technology will remain and reach a homeostasis at the level where it’s actually a helpful tool.
When the dotcom bubble popped the internet didn’t disappear, it just became what it was supposed to be. When the AI bubble pops the technology will still be there for use cases where it’s actually appropriate and useful.
I think the AI bubble will pop when another part of the economy needs capital, sucking up the capital going to AI.
AI burning money for market share isn’t new to Wall Street, that’s been tech for the past generation. The issue is going to be liquidity to not test these valuations. Once they are tested, they are going to collapse.
Dogs and cats living together etc.
Social media companies will fill the gaps, continuing to spread their toxicity through society unchecked. The Pied Piper of Hamelin is starting to look more like a cautionary tale.
One change for me is celebrating on the anniversary of an AI company shutting down.
Big tech will keep their AI. Like others here have said, AI companies without another income stream will go belly up.
Some data centers will be sold and repurposed. Others will sell off their hardware to the consumer market.
Lower temperatures and less noise near the former data centers.
I don’t see property values going back to where they were unless a data center is dismantled and the water table recovers.
