In the spring of 2026, Nasdaq overhauled its index rules to eliminate the traditional three month seasoning period and minimum float requirements, allowing massive IPOs like SpaceX to enter the Nasdaq 100 after just 15 trading days. This forced passive funds and retail retirement accounts to absorb over valued companies before proper price discovery.
Simultaneously, stricter listing standards raised public float and SPAC thresholds, choking off smaller, high growth IPOs. Together, these changes concentrated the exchange in mega cap tech and AI names without historical safeguards, leaving the index uniquely vulnerable when AI sentiment soured, culminating in a Nov. sell off triggered by Anthropic's post IPO plunge and Nvidia's regulatory turmoil that wiped out the Nasdaq's yearly gains.
Yep. Nasdaq got greedy. For their greed they’re gonna lose market trust.
Before a big valuation meant something. Now a big IPO is a Trojan horse to steal retirement money and HSA money from the index investors in the hands of VCs.
Look at Uber, AirBnb, DoorDash? The companies come later to market when the growth is all captured. Then dump on retail.
I wonder where the floor is. AI does have SOME value, even if it's not nearly what the industry claims it is. What is it about the mid 2020s that has Americans so detached from reality?
Let's just ignore all the social/moral/political issues with AI, and focus on why companies like Anthropic are absolutely screwed.
The first problem is that AI companies have no way of keeping their clients captive. It's not like conventional software, because clients can just switch to another provider in a day, with minimal disruption. Which means that they have ZERO guarantee of future business if something cheaper, better, or even more socially acceptable comes out in the next few years.
Secondly, companies like Anthropic made commitments to purchase compute/chips at the worst possible time. Basically, they overpaid, and any competition that comes along after the bubble pops will be able to buy up unwanted data centers (or compute) for pennies on the dollar. Which makes Anthropic inherently uncompetitive in terms of profitability.
Third, there's absolutely nothing to prevent other companies from stealing their stuff, slapping another name on it, and calling it good. Because nobody using AI at scale cares about other people's intellectual property. Which means that all those billions spent on training/research might as well have been flushed down the toilet.
TLDR: They've got no customer loyalty, they're at a financial disadvantage, and there's nothing stopping people from stealing their work. So, they might limp on for a bit, but they're screwed in the long term. And it's only a matter of time before someone else eats their lunch... without having to pay for it.
It's also incredibly difficult to regulate, and there's no reasonable argument for why these initial companies -- who keep repeatedly proving themselves uninterested in or incapable of governance -- should be carved out any exceptions to the rules they now claim we need.
Well written. Basically, these companies desperately convinced themselves that they had enormous first-mover advantage, and they simply don't. There are some people and companies of course who will become "hooked" on specific models, but most won't care and will use whatever is cheapest and most convenient.
Switching is easy, like you say. And ironically, it's because these models are so easy to use, which is one of their selling points.
I expect MSFT will ultimately emerge as the winner here. Office 365 is deployed at almost every Fortune 500. They are great at workflow integration, and Copilot’s GUI and app interfaces will get better and better. They’ll route most work to open-weight models, which will be good enough for nearly all tasks. And compute will be just another utility, but without the geofenced monopolies that allow utilities to maintain pricing, and the ever-dangerous business mix of high fixed cost and low marginal cost. All that will result in rapid token price erosion. Corporate customers will get big productivity gains and MSFT will further strengthen its Office moat and subscription revenue. Infrastructure and models will be a race to the bottom.
Microsoft is widely used, yes, but their products lag behind their competitors (copilot vs any other AI, even Microsoft word/excel vs. Google). I work at a decent-sized corporate company that is now going hard with Claude. However, a big benefit to Claude is that it is compatible with the Microsoft ecosystem.
Opus5.5 is available within Copilot. So are GPT Astra and Sol.
I also use Claude right now, but I can already see where Microsoft is going. Their GUI will be better (more visual studio like approach to organizing projects, vs. Claude’s very limited pinned system). Their office integration will be better because they control those products. The security will be easier to manage because Microsoft controls windows, OneDrive, etc.
And they’ll just auto-route the work to different models depending on the task, allowing them to have both the most adept open weight models as well whatever frontier model is 6 months ahead at that moment in time. And the corporate customer just has one company to work with and one budget to manage.
Their entire business is literally just a few models. Other models have caught up. In several benchmarks open weight is even surpassing them. They're screwed.
At least with Google, Microsoft, Amazon, etc.. the models aren't their entire business. They own hardware and have other revenue streams. I think every company like Anthropic that didn't diversify their business will all go under. Guaranteed. I'm betting Anthropic gets bought by Google or some other megacorp after the stock tanks into oblivion.
>The first problem is that AI companies have no way of keeping their clients captive. It's not like conventional software, because clients can just switch to another provider in a day, with minimal disruption. Which means that they have ZERO guarantee of future business if something cheaper, better, or even more socially acceptable comes out in the next few years.
This isn't true. AI companies make their money the same way Amazon makes its money. It is a middle man between the business, and the customer which facilitates the entire process. For a business to provide its service at a better margin due to increased efficiency and market competitiveness demands resulting from AI, it has to use an AI model. That business is now trapped by a licensing agreement, and subscription the same way for a retailed to function it simply has to concede to Amazon, and pay whatever fees it has to pay.
The issue is AI has yet to be applied at a scale where the subscription model will work. This is where the value questions emerge and this is where our attention should be directed. The underlying model is already there, and it will make money, so the question is "why hasn't AI been adopted as anticipated?". So far, AI is being used by the general population for making videos about Lil Vader, as an aid to help young children learn math, write low effort automatic emails and to write low-effort, and factually incorrect essays, articles and papers. None of these uses had a particularly high value pre-AI, so it's hard to see why AI would suddenly generate value in these contexts?
I think its because AI is being sold as an all powerful, all applicable tool for any, and all tasks, and organizations when in reality it is probably hyper niche and better suited for dealing with research and security/defense. I think the answer to the "why?" question above is "because AI isn't actually that useful, or worth the headache for a majority of businesses". In the commercial sector, it seems to be the norm that most major, or large firms are reporting that AI doesn't see to me as useful as they thought. The big winners from AI are the small SMEs who run their business out of the living room, or as a sole trader who can use AI to help them handle tasks, and paperwork they would otherwise have to hire someone to handle. These small SMEs are simply not large enough firms, nor economic actors to support the kind of funding AI companies need.
You can't fund the single largest allocation of capital in human history, and subsidize the entire US economy as its national debt doubles with AI brain rot, and slightly more advanced automated responses to emails and queries. I've paid for both ChatGPT and Gemini. Neither can even generate graphs using data I've hand collected and notated. I had to do it by hand because they can't collect the data even with carefully picked parameters, key words and methodology. The fuck are we even doing here? It's legit easier, and quicker for me to just make the graph by hand rather than spend 2 hours feeding prompts where it seems to fuck up stuff it already got right before resulting in a never ending cycle of deeper confusion and hallucination. AI can't even provide reliable answers to simple questions half the time. Shit, reddit is the fucking source half the time. If the US is expecting reddit comments to save its economy then fuck me, we're all doomed.
OK, at your first point, you make the argument that a walled garden is necessary to keep customers in, by hampering migration to a new service. I think this isn't necessary if you keep making a good product, and keeping a walled garden approach is a critical factor for overall enshittification, as it allows for worsening service without making people migrate.
I'm not disagreeing with your other points, but I wanted to address this as I am not convinced it should be normalized as good business practice.
No. I never said that walled gardens were "necessary" for customer retention. I just pointed out that Anthropic doesn't have one.
Which really sucks for them, and doesn't bode well for their financial future. Because they spent a lot of money, and racked up a ton of debt to create a product that will inevitably be stolen, or copied. And being that far in debt makes it very hard for them to compete on an equal footing with a company that isn't spending billions on interest payments.
Though, I'm 100% with you on walled gardens being terrible for consumers, and I wish they were legislated out of existence.
The floor is pretty low. There's enough leverage around to just wipe out some industries. Not that we will actually see that of course. But there is hardly a natural firebreak
AI will stay for sure, Anthropics maybe will, maybe wont. The bubble inflated prices and revenues, after the bubble demand and price will drop sharply and then level and rise again, but we don't know who will be there and who wont.
From article by Lauren Almeida and Graeme Wearden:
As Anthropic’s stock dropped during its second day of trading, the wider stock market dipped with it. On Friday, the Nasdaq Composite erased all its gains for 2026 and closed down 5% from its value on 1 January. The Dow Jones Industrial Average and S&P 500 closed in correction territory.
Anthropic’s record-breaking IPO failed to stem a global sell-off in AI-linked companies, and the tech-heavy Nasdaq erased some $4tn in market value this week, according to the Wall Street Journal.
Shares in the frontier AI company behind the Claude chatbot, which floated in the US on Thursday, fell by 9% on Friday, ceding most of the gains from the biggest initial public offering in history. They dropped to $88.30, towards the $83 level at which Anthropic priced its stock market debut.
Anthropic’s first two days as a listed company have been volatile. Its shares jumped by 17% on Thursday, after briefly surging by almost 40% when trading in the firm began.
US technology stocks followed Anthropic down on Friday, with the memory chipmaker SanDisk falling 14% and rival Micron down 12%. Nvidia fell 14%, eliminating hundreds of billions of dollars in market cap.
Neil Wilson, a UK investor strategist at Saxo, said there was “just relentless selling in tech”. He warned: “Anthropic shares tumbled to take an $85 handle as the post-IPO euphoria gives way to the reality of a steep sell-off across the tech and AI space that shows no sign of easing just yet.”
Global markets slumped on Friday too; Asian stock markets recorded their worst week since 2020, as shares in big chip companies fell. In Taiwan, stocks dropped 6.7% to hit their lowest point since mid-April, with the chip maker TSMC down about 6%.
DiligentAd7956 | 9 hours ago
In the spring of 2026, Nasdaq overhauled its index rules to eliminate the traditional three month seasoning period and minimum float requirements, allowing massive IPOs like SpaceX to enter the Nasdaq 100 after just 15 trading days. This forced passive funds and retail retirement accounts to absorb over valued companies before proper price discovery.
Simultaneously, stricter listing standards raised public float and SPAC thresholds, choking off smaller, high growth IPOs. Together, these changes concentrated the exchange in mega cap tech and AI names without historical safeguards, leaving the index uniquely vulnerable when AI sentiment soured, culminating in a Nov. sell off triggered by Anthropic's post IPO plunge and Nvidia's regulatory turmoil that wiped out the Nasdaq's yearly gains.
african_cheetah | 7 hours ago
Yep. Nasdaq got greedy. For their greed they’re gonna lose market trust.
Before a big valuation meant something. Now a big IPO is a Trojan horse to steal retirement money and HSA money from the index investors in the hands of VCs.
Look at Uber, AirBnb, DoorDash? The companies come later to market when the growth is all captured. Then dump on retail.
BenL90 | 13 minutes ago
Remember NASDAQ now isn't Financial company anymore based on what they think. It's a Technology company.. so that's why I think so
ICLazeru | 14 hours ago
I wonder where the floor is. AI does have SOME value, even if it's not nearly what the industry claims it is. What is it about the mid 2020s that has Americans so detached from reality?
TheDeliciousMeats | 12 hours ago
Let's just ignore all the social/moral/political issues with AI, and focus on why companies like Anthropic are absolutely screwed.
The first problem is that AI companies have no way of keeping their clients captive. It's not like conventional software, because clients can just switch to another provider in a day, with minimal disruption. Which means that they have ZERO guarantee of future business if something cheaper, better, or even more socially acceptable comes out in the next few years.
Secondly, companies like Anthropic made commitments to purchase compute/chips at the worst possible time. Basically, they overpaid, and any competition that comes along after the bubble pops will be able to buy up unwanted data centers (or compute) for pennies on the dollar. Which makes Anthropic inherently uncompetitive in terms of profitability.
Third, there's absolutely nothing to prevent other companies from stealing their stuff, slapping another name on it, and calling it good. Because nobody using AI at scale cares about other people's intellectual property. Which means that all those billions spent on training/research might as well have been flushed down the toilet.
TLDR: They've got no customer loyalty, they're at a financial disadvantage, and there's nothing stopping people from stealing their work. So, they might limp on for a bit, but they're screwed in the long term. And it's only a matter of time before someone else eats their lunch... without having to pay for it.
badnewzero | 10 hours ago
Which is why they keep harping on about how AI is dangerous, so that governmental regulation can become their moat
TheDeliciousMeats | 10 hours ago
Which sounds like it could work, except for one minor problem: politicians are only as loyal as their last paycheck.
The second companies like Anthropic start running out of cash, they're done for. No more favorable regulation, no more moat.
Qeltar_ | 6 hours ago
It's also incredibly difficult to regulate, and there's no reasonable argument for why these initial companies -- who keep repeatedly proving themselves uninterested in or incapable of governance -- should be carved out any exceptions to the rules they now claim we need.
badnewzero | 2 hours ago
Oh yeah totally agree. Not saying it’s a good play, it just seems to be their only play
Qeltar_ | 6 hours ago
Well written. Basically, these companies desperately convinced themselves that they had enormous first-mover advantage, and they simply don't. There are some people and companies of course who will become "hooked" on specific models, but most won't care and will use whatever is cheapest and most convenient.
Switching is easy, like you say. And ironically, it's because these models are so easy to use, which is one of their selling points.
PIKFYVE | 4 hours ago
I expect MSFT will ultimately emerge as the winner here. Office 365 is deployed at almost every Fortune 500. They are great at workflow integration, and Copilot’s GUI and app interfaces will get better and better. They’ll route most work to open-weight models, which will be good enough for nearly all tasks. And compute will be just another utility, but without the geofenced monopolies that allow utilities to maintain pricing, and the ever-dangerous business mix of high fixed cost and low marginal cost. All that will result in rapid token price erosion. Corporate customers will get big productivity gains and MSFT will further strengthen its Office moat and subscription revenue. Infrastructure and models will be a race to the bottom.
goonie814 | 3 hours ago
Microsoft is widely used, yes, but their products lag behind their competitors (copilot vs any other AI, even Microsoft word/excel vs. Google). I work at a decent-sized corporate company that is now going hard with Claude. However, a big benefit to Claude is that it is compatible with the Microsoft ecosystem.
PIKFYVE | 2 hours ago
Opus5.5 is available within Copilot. So are GPT Astra and Sol.
I also use Claude right now, but I can already see where Microsoft is going. Their GUI will be better (more visual studio like approach to organizing projects, vs. Claude’s very limited pinned system). Their office integration will be better because they control those products. The security will be easier to manage because Microsoft controls windows, OneDrive, etc.
And they’ll just auto-route the work to different models depending on the task, allowing them to have both the most adept open weight models as well whatever frontier model is 6 months ahead at that moment in time. And the corporate customer just has one company to work with and one budget to manage.
krileon | 4 hours ago
Their entire business is literally just a few models. Other models have caught up. In several benchmarks open weight is even surpassing them. They're screwed.
At least with Google, Microsoft, Amazon, etc.. the models aren't their entire business. They own hardware and have other revenue streams. I think every company like Anthropic that didn't diversify their business will all go under. Guaranteed. I'm betting Anthropic gets bought by Google or some other megacorp after the stock tanks into oblivion.
agumonkey | 6 hours ago
would be weird if 2026 AI would end up dilluted by economic friction and perverse effects
Thom0 | 5 hours ago
>The first problem is that AI companies have no way of keeping their clients captive. It's not like conventional software, because clients can just switch to another provider in a day, with minimal disruption. Which means that they have ZERO guarantee of future business if something cheaper, better, or even more socially acceptable comes out in the next few years.
This isn't true. AI companies make their money the same way Amazon makes its money. It is a middle man between the business, and the customer which facilitates the entire process. For a business to provide its service at a better margin due to increased efficiency and market competitiveness demands resulting from AI, it has to use an AI model. That business is now trapped by a licensing agreement, and subscription the same way for a retailed to function it simply has to concede to Amazon, and pay whatever fees it has to pay.
The issue is AI has yet to be applied at a scale where the subscription model will work. This is where the value questions emerge and this is where our attention should be directed. The underlying model is already there, and it will make money, so the question is "why hasn't AI been adopted as anticipated?". So far, AI is being used by the general population for making videos about Lil Vader, as an aid to help young children learn math, write low effort automatic emails and to write low-effort, and factually incorrect essays, articles and papers. None of these uses had a particularly high value pre-AI, so it's hard to see why AI would suddenly generate value in these contexts?
I think its because AI is being sold as an all powerful, all applicable tool for any, and all tasks, and organizations when in reality it is probably hyper niche and better suited for dealing with research and security/defense. I think the answer to the "why?" question above is "because AI isn't actually that useful, or worth the headache for a majority of businesses". In the commercial sector, it seems to be the norm that most major, or large firms are reporting that AI doesn't see to me as useful as they thought. The big winners from AI are the small SMEs who run their business out of the living room, or as a sole trader who can use AI to help them handle tasks, and paperwork they would otherwise have to hire someone to handle. These small SMEs are simply not large enough firms, nor economic actors to support the kind of funding AI companies need.
You can't fund the single largest allocation of capital in human history, and subsidize the entire US economy as its national debt doubles with AI brain rot, and slightly more advanced automated responses to emails and queries. I've paid for both ChatGPT and Gemini. Neither can even generate graphs using data I've hand collected and notated. I had to do it by hand because they can't collect the data even with carefully picked parameters, key words and methodology. The fuck are we even doing here? It's legit easier, and quicker for me to just make the graph by hand rather than spend 2 hours feeding prompts where it seems to fuck up stuff it already got right before resulting in a never ending cycle of deeper confusion and hallucination. AI can't even provide reliable answers to simple questions half the time. Shit, reddit is the fucking source half the time. If the US is expecting reddit comments to save its economy then fuck me, we're all doomed.
nucleardk | 8 hours ago
OK, at your first point, you make the argument that a walled garden is necessary to keep customers in, by hampering migration to a new service. I think this isn't necessary if you keep making a good product, and keeping a walled garden approach is a critical factor for overall enshittification, as it allows for worsening service without making people migrate.
I'm not disagreeing with your other points, but I wanted to address this as I am not convinced it should be normalized as good business practice.
TheDeliciousMeats | 7 hours ago
No. I never said that walled gardens were "necessary" for customer retention. I just pointed out that Anthropic doesn't have one.
Which really sucks for them, and doesn't bode well for their financial future. Because they spent a lot of money, and racked up a ton of debt to create a product that will inevitably be stolen, or copied. And being that far in debt makes it very hard for them to compete on an equal footing with a company that isn't spending billions on interest payments.
Though, I'm 100% with you on walled gardens being terrible for consumers, and I wish they were legislated out of existence.
MaximumFamiliar9080 | 14 hours ago
Ai has some value but not relative to its cost.
Its possibly valuable but not profitable.
BoDrax | 4 hours ago
>What is it about the mid 2020s that has Americans so detached from reality?
>>The unregulated social media algorithms and having to fit worldviews into a world that doesn't match them.
Qwertycrackers | 3 hours ago
The floor is pretty low. There's enough leverage around to just wipe out some industries. Not that we will actually see that of course. But there is hardly a natural firebreak
zanzara1968 | 3 hours ago
AI will stay for sure, Anthropics maybe will, maybe wont. The bubble inflated prices and revenues, after the bubble demand and price will drop sharply and then level and rise again, but we don't know who will be there and who wont.
DinoDonkeyDoodle | 2 hours ago
All I know is the time to buy into ai is coming. Those who survive the pop, I mean.
Dismal_Buy3580 | 10 hours ago
The floor????
My guy this is going to be a correction that keeps on giving.
mtbdork | 13 hours ago
lol, no it doesn’t.
[OP] marketrent | 18 hours ago
From article by Lauren Almeida and Graeme Wearden:
As Anthropic’s stock dropped during its second day of trading, the wider stock market dipped with it. On Friday, the Nasdaq Composite erased all its gains for 2026 and closed down 5% from its value on 1 January. The Dow Jones Industrial Average and S&P 500 closed in correction territory.
Anthropic’s record-breaking IPO failed to stem a global sell-off in AI-linked companies, and the tech-heavy Nasdaq erased some $4tn in market value this week, according to the Wall Street Journal.
Shares in the frontier AI company behind the Claude chatbot, which floated in the US on Thursday, fell by 9% on Friday, ceding most of the gains from the biggest initial public offering in history. They dropped to $88.30, towards the $83 level at which Anthropic priced its stock market debut.
Anthropic’s first two days as a listed company have been volatile. Its shares jumped by 17% on Thursday, after briefly surging by almost 40% when trading in the firm began.
US technology stocks followed Anthropic down on Friday, with the memory chipmaker SanDisk falling 14% and rival Micron down 12%. Nvidia fell 14%, eliminating hundreds of billions of dollars in market cap.
Neil Wilson, a UK investor strategist at Saxo, said there was “just relentless selling in tech”. He warned: “Anthropic shares tumbled to take an $85 handle as the post-IPO euphoria gives way to the reality of a steep sell-off across the tech and AI space that shows no sign of easing just yet.”
Global markets slumped on Friday too; Asian stock markets recorded their worst week since 2020, as shares in big chip companies fell. In Taiwan, stocks dropped 6.7% to hit their lowest point since mid-April, with the chip maker TSMC down about 6%.
BenL90 | 16 hours ago
Amateur. Indonesian Stock Market lost 50% of it's value because of dual brother and it's sidekicks.