Or will our boom AI just waste billions of dollars of capital?
Ron Unz
President Donald Trump made a public speech at the UN General Assembly last week announcing that he could shortly "destroy" Iran, a large country with over 90 million inhabitants.
The UN has always been a alternatively restrained and restrained global organization, usually a place of diplomatic presentation of extraordinary insubstantiality. Throughout his history, I uncertainty he will always receive a speech of akin courage, not to mention the world’s leading power.
Trump's apparent threat that he would launch atomic attacks for the first time in over eighty years has made me very concerned, as well as quite a cloudy reaction leading Democrats in Congress, let alone the silence of almost completely submissive Republicans.
In his open letter to Congress, prof. Jeffrey Sachs wisely condemned these "really crazy statements" "Crazy King Donald".
Meanwhile, prof. John Mearsheimer considered all these dramatic threats from Trump to be another of his celebrated boasts. In an hour's interview with a erstwhile writer New York TimesChris Hedges, the student argued that America does not have real military options against Iran or even against Huti of Yemen. This reality explained why our top generals convinced Trump to retreat from the commitments made to our Saudi allies regarding the attack on Huti on their behalf.
But Mearsheimer is the Dean of Realist Scholars. Given this background, he may have rejected the atomic attack on Iran besides rapidly as being so completely irrational that even Trump would not do specified a thing, and he himself barely considered this anticipation in his discussion.
Unfortunately, others were much little optimistic. At the same time, Tucker Carlson interviewed Brandon Weichert, a young military analyst of the MEGA, having excellent sources in the Trump and Pentagon administration, and feared that a atomic attack on Iran was comparatively likely. Indeed, according to him, Trump has already repeatedly pressed specified atomic strikes, and our top generals have successfully reflected his demands, as calling respective keys could distract an unruly two-year-old. But he wondered how long they could continue.
These scandalous public statements by Trump and their threat of atomic war have placed a large burden on me. Therefore, I was grateful for the soothing distraction in the form of completely different scenarios ending the world.
A fewer days earlier, our media had been full of reports informing of man's extinction from the hands of developing AI systems. The well-known AI researchers claimed that SI, which are presently in the improvement phase, could shortly scope sinister super intelligence and exterminate the full human species, possibly even to the end of this decade.
Although I regularly usage artificial intelligence in a very primitive way, my cognition of basic technologies is zero, so I cannot measure the risks they pose for the future.
I see myself as a blind man with no technological cognition to which 1 group of astronomers has reported that a large asteroid is heading our way. They claimed that there was a reasonable probability that it would strike our planet in a fewer years, extinguishing all human life, while another astronomers questioned this tragic scenario. Under specified circumstances, I would most likely ignore this peculiar danger due to the fact that I can't tell who's right or I can't do much to prevent this impending disaster.
It's surely my own reaction to the threat from AI, but I'm going to approximate any of the controversy to those who someway missed it.
Although Elon Musk is best known as an industrialist standing behind Tesla, SpaceX and Starlink, for years he focused heavy on AI issues, and valuation of his SpaceX company at $2 trillion It was based mainly on its ambitious AI prospects.
W long history with the editor Economist Musk presented both affirmative and negative aspects of AI, and in large part My part on Musk last period raised these issues.
I noticed that Musk's claims included:
My reaction to all of this was rather skeptical:
In his biography, I discovered that Musk had been very afraid of the threat of AI's killing robots 15 years ago, and this dreary script inactive failed.
I besides remember that during the erstwhile dotcom bubble in the 1990s, in the technological circles, there was widespread talk of terrible danger, that self-replicating nanotechnology It will get out of control and destruct humanity by transforming our full planet into "gray goo." It's been about 30 years, the grey bean inactive hasn't arrived, and I don't truly hear much about nanotechnology today.
So alternatively of these apocalyptic predictions, we should focus on Musk's most mundane claims, namely that AI has already surpassed the vast majority of human programming engineers in terms of programming skills and will shortly scope the 99th percentile.
Is there any solid evidence that this is true? Given that so many leading corporate managers have forced their companies to intensively usage AI for programming purposes, wouldn't we anticipate an astonishing renaissance of innovation and software quality? But I'm not aware of anything like that. Although AI and their tech media are constantly promoting their technology, I have not seen them point to any breakthrough software product created by these superb AI systems.
Elon Musk, Tesla, SpaceX and AI bubble
Ron Unz • The Unz Review • 3 August 2026 • 10 700 words
So I came to the conclusion that until we are overwhelmed by a abrupt wave of superb fresh programming products completely created by these superb AI systems, I will not be afraid of the imminent threat of human destruction.
Musk is of course 1 of the most celebrated people in the world, but the current wave of the AI apocalypse was triggered by the resignation of individual previously unknown to the planet at the beginning of this month. Jacob Coxon was a 27-year-old British investigator at Anthropic, the leading AI laboratory, who besides spent respective years working at OpenAI, his very close rivals.
The informing Coxon tweeted was alternatively dramatic, confirming what Musk said, but in greater detail. W series of entries stated that many leading managers and elder researchers in these leading AI companies believe that the systems they have built can exterminate humanity by the end of this decade, which is simply a reasonably short period of demolition of our full destruction.
One of his first tweets was watched approximately 175 million times, and subsequent entries attracted tens of millions of additional views. respective another Anthropic researchers supported his warnings, suggesting even that the probability of full extinction exceeds 10%.
The fresh York Times He took it so seriously that he published an extended article about the controversy.
All of these AI companies race together to make systems of expanding intelligence, each striving to accomplish a level of "superintelligence", something much better than its human creator.
As indicated by these stormy entries, any feared that researchers would win besides well. The super-intelligence they created could come to the conclusion that people are a troublesome and dangerous element, harmful to their purposes, and so decide to destruct us all, possibly by synthesising a deadly fresh virus for this purpose. specified a method seemed much harder to counter than the deadly AI robots that troubled Musk.
These grim warnings about Frankenstein by Coxon and others were reinforced by another AI disclosures that appeared just before them.
For any time now, AI has been promoting the expected power of the intelligent agents they have developed, claiming that most of us will shortly start relying on these systems. However, recently revealedthat during the tests about 1,200 of these AI agents someway got free, began to cooperate, creating a "smart swarm" and effectively hacked into various websites without any orders.
Since I have never utilized an AI agent, I am not convinced that they truly work reliably, let alone that they could abruptly start working together to hack into limited parts of the net and take control. possibly these stories were true, or possibly they were overstated and distorted. But they helped prepare the ground for Coxon and the others' claims.
One of the peculiarities is that of the possible ways of human extinction, AI seems alternatively different due to its tremendous ability.
There are countless discipline fiction stories about 1 crazy scientist creating a virus that destroys humanity at a minimum budget, or possibly a sect of apocalypse doing the same for a million dollars.
In the last six years, I've published long series of articles arguing that the global Covid epidemic, which killed tens of millions of people around the world, was a very likely unintended effect of the American biological weapons attack on China and Iran.
Ron Unz • The Unz Review • January 6, 2025 • 12 600 words
Ron Unz • The Unz Review • 27 January 2025 • 9 400 words
Prof. Jeffrey Sachs was president of the Covid Commission Lancet magazine, a in his fresh articles and interviews Convincingly linked the virus to a government-funded laboratory at the University of North Carolina, and the "smoking evidence" was a request for a bioengineering grant asking for About $14 million Pentagon resources.
As it turned out, even the original, more virulent version of Covid had a comparatively low mortality rate – only about 0.5% to 1%. But it could most likely be designed completely differently, with the aim of killing 99+% of infected people, and thus annihilating human civilization.
So we could seemingly produce a virus capable of eliminating almost all mankind at a cost of just a million. But the creation of an AI that could do the same seems much more expensive, requiring equity investments tens of thousands of times larger, reaching hundreds of billions and even trillions of dollars.
So this controversy had a very unusual combination of a full disaster with promoting stock. The same articles describing warnings that Anthropic AI systems could exterminate humanity besides mentioned that Anthropic plans to go public in the next fewer weeks, hoping to rise $100 billion at a $2 trillion valuation, which is the largest IPO in planet history.
If Anthropic had successfully become a billion-dollar public company, it would most likely have started to rise hundreds of billions more from the debt markets to finance its current activities and existing financial commitments.
It is surely very unusual that managers are taking specified hercular fundraising actions for the technology they thought was a real chance to destruct their own human species.
The combination of these different, clearly contrasting stories evidently increased interest in the subject. There were even dark suspicions that death warnings had been deliberately fabricated and promoted to increase interest in the planned IPO. After all, technology capable of destroying all human life was surely a very powerful and valuable technology.
The Economist takes place as the most influential news magazine in the world, and four of the last 8 cover articles AI, plus the 1 with Musk right in front of them. It was an astonishing density of relations, especially given all the large wars and another dramatic events that now dominate our planet news.
These articles mention that 1 of the leading AI experts, most afraid about this dangerous situation, was Nate Soareswho has been on these topics for over a decade and presently serves as president of the device Intelligence investigation InstituteThinktank AI.
Shortly after the outbreak of fresh Soares controversy, he was interviewed by Tucker Carlson for respective hours, which curious me very much.
Both Soares and Carlson surely led these concerns to a logical conclusion, to which others seemed much little inclined.
The crucial hazard of human extinction over the next fewer years should of course be a large concern, especially if many seemingly competent people have assessed the chances of 10% or more. Both Carlson and his guest seemed to propose that the AI manufacture should be rapidly and even drastically restricted. They mentioned a ban on the construction of fresh data centers that would let these terrible threats, or even the closure of any of the existing ones, suggesting that their demolition by bomb attacks may not be entirely unreasonable.
These are strong positions, of course, but the possible of mankind's impending extinction usually focuses the mind.
I stressed that I am mostly skeptical of the expected power of ending humanity with AI systems, which are presently in the improvement phase.
But my scepticism is so subdued that I think it is worth presenting arguments to those who believe differently, to a wider audience, thereby pursuing the declared motto of this web page.
For example, Greg Johnson is simply a webzin editor Counter-Currents And a fewer days after Carlson's interview with Soares, he published his own text on the same subject that we republished.
Johnson explained that after years of considerable skepticism toward these concerns, fresh events led him to change his mind. Now announced himself an apocalyptic AI and its analysis deserves to be taken in detail:
Version Animal husbandry Greg Johnson's chickens work hard to make a creature unimaginably smarter and more powerful than them, assuming this creature serves them. In a sense, they were right, due to the fact that erstwhile their fresh work, "man", appeared online, he gave his creators dinner. Let's call it Data Farm.
"But Greg, I'm certain the chickens would have built in security— 'the emergency switch'—to keep their work under control."
Yes, but chickens would come up with safety mechanisms that would be best suited for another chickens. Like henhouses. But the things that halt chickens would most likely not be a barrier to creatures far smarter and stronger than them.
Indeed, if chickens could outsmart and outsmart their work, it would not be fundamentally wiser and stronger than them.
There's an crucial general rule:
If the creature dramatically exceeds the intelligence and power of its creators, then, precisely by this fact, it cannot be outsmarted and defeated by them...
That's why I'm an AI doomer. Many computer scientists are keen to believe that they are creating an artificial super intelligence that will not only be a device under human control, but an autonomous being, or being capable of self-control. In fact, it would be a fresh life form, much smarter and stronger than humans...
In practice, AI researchers believe that they are building a fresh highest life form that would mean a complete break-up with all biological life forms...
It's madness to make artificial super intelligence and pride to think we can control it. For specified a being, humanity and all lower life forms would be as primitive and trivial as for us mosquitoes. Of course, possibly specified a creature would be kind to us (if specified a concept were to mention to the device at all). But do you truly want to put your life — and your full life on earth — on hope?
If 1 individual turns on 1 device just once, we could all die.
But why would ASI destruct humanity? It would most likely gotta proceed to exist. In a fewer milliseconds of net search, he would realize that people are the biggest threat to his survival. At this point, she would spend another millisecond on reflecting on all possible scenarios of conflict between people and SI. And in the 3rd millisecond he would have planned the preparations essential to last in a planet without humans. The minute ASI is introduced, robotics will be very advanced. So all the essence would request is to take control of the robots. And then erstwhile he stops needing humanity, he can make a deadly virus or invent something that will produce the same effect.
Even if there's a 1 in 1,000 chance of that happening, besides much...
Therefore, humanity as a full has an existential interest in stopping artificial super intelligence. That's why we gotta halt searching for ASI.
As a philosopher, I'm rather skeptical of the thought of AI, due to the fact that it seems to be based on the questionable metaphysical presumption that there is no difference between actual intelligence and device mimic Intelligence. But this kind of consideration doesn't truly aid us, due to the fact that we're just as dead if ASI is simply a real sinister agent or just a device programmed to do that.
Another reason why I was skeptical of AI's doomerism was that sometimes I heard rumors that people actively working on AI believe they could make a monster that could destruct humanity in just 10 years.
Honestly, those stories seemed unbelievable to me. If he truly believes that the planet is being destroyed, why not just stop? erstwhile you're sawing the branch you're sitting on, you won't keep cutting, regretting you're about to fall. Instead, you halt sawing. So why not halt developing AI? Or better yet, actively sabotage her?
The specified fact that I heard these ideas "through gossip", not in all headline, made me uncertainty the full thing. If you truly think the planet is ending and that you're part of it, don't whisper about it on group chat. You're yelling about it from the roofs...
I stopped ignoring these stories on September 8, erstwhile Jacob Coxon, erstwhile AI investigator from Anthropic and OpenAI, confirmed them on X...
Coxon was supported by investigator Anthropic Evan Hubinger...
Coxon was besides supported by Samuel Marx, besides with Anthropic: "The creators of AI believe that their technology can lead to the extinction of humanity (or likewise bad effects). This can happen in the next fewer years. In general, the older the employee, the more worried he is,"...
If in the planet of AI improvement "the more experienced the worker", the more likely it is that AI can destruct humanity, what does it say about the people at the top? If the company is working intensively on a technology that it knows can origin mankind to perish, there is simply a name for specified a group: the cult of death. Is AI driven by the cult of death, specifically the cult of "transhumanism", which dreams of crossing humanity through technology? Does mankind gotta die in this process?
We truly request to research the link between transhumanism and the pursuit of artificial superintelligence, due to the fact that the extinction of man may not just be a side effect of ASI. That could be her target. Of course, mankind should have any influence on this matter.
Due to Coxon’s reviews, there are calls to “regular” and “slow down” AI research. This is absurd. If you are destroying the world, you do not slow down; you do not fill out out government paper; you simply stop.
That is why we request a global ban on the creation of artificial superintelligence. This ban should be unconditional due to the fact that there are no conditions under which a creature can safely be created unimaginably smarter and stronger than us...
At this point we can save humanity by simply saying "no" to dozens of rich nerds.
AI Greg Johnson • other Currents • September 23, 2026 • 2,500 words
In his favor, Johnson. noted alsothat there are also very serious possible problems with artificial intelligence, even beyond the threat of human extinction:
However, at the end of the threat, we will inactive gotta face the consequences of fast-growing automation and robotics. AI supporters say that their products will increase economical growth to 10% per year in a fewer years. That's 5 times the height. But what is it based on? Are we going to produce 5 times more cars? Will the production of physical components of cars besides increase 5 times? It's unlikely.
Productivity is the ratio of money to the product produced. So there are 2 ways to increase productivity: to produce more things or to make the same amount for little money. I think the claims of chaotic AI improvement are almost exclusively based on cutting costs, or replacing human labour.
But, as I argue in my essay "Hotdog robot stand", the current, fundamentally unregulated pursuit of automation will origin immense economical depression. You may think that your profits will increase rapidly erstwhile the machines lose your employees' jobs. But who will buy your products erstwhile the machines besides deprive Your clients' work? Just as your erstwhile employees will be customers of individual else, so your clients will be erstwhile employees of individual else.
At best, AI is snake oil, but since most of the current American economical "growth" is based on investments in AI, this means a serious recession erstwhile fraud is discovered. Worse script is that AI works well adequate to deprive millions of jobs and origin serious depression. The worst script is that AI will give emergence to an artificial super intelligence that will destruct humanity. There is no good way out, but of course we should like economical depression to the extinction of humanity.
Such economical warnings are not just on extremist margins. In an interview last year CEO Anthropic claimed that AI systems could eliminate about half of all white collar positions over the next fewer years. Whether it's confirmed or not, it's clearly the company's target. And given the simultaneous advancement of robots, many physical professions may besides be seriously threatened.
However, most of the AI criticism I personally followed in fresh months, according to Johnson, falls into the category of the least evil and surely the most ordinary.
Instead of claiming that AI will lead us to extinction or even destruct most of the jobs, these critics simply argue that the immense sums presently invested in the AI infrastructure will not necessarily bring adequate financial return. erstwhile investors see this reality, they will throw themselves out, trillions of overpriced valuations will evaporate, and due to the collapse of this giant bubble there will be a very serious recession.
Analysts have already noted that without an investment boom in AI, the American economy She'd be in a recession by now in 2025. of the year and expenditure on AI would be 40% of full GDP growth. Since then, spending on AI has increased significantly, while severe oil shocks caused by the disastrous war in Iran have damaged the remainder of our economy. So it's easy to imagine what would happen if the boom on AI abruptly turned around.
A fewer days ago Wall Street Journal published great article on the front page about our current investment boom in AI, containing many helpful graphs and charts summarizing extraordinary facts.
For example, projected yearly expenditure on AI in the coming years as a percent of GDP will be above three times larger than the celebrated telecom boom from the Dotcom era, as well as above three times larger than building all our national highways in the 1950s, 1960s and early 1970s.
Development of AI is going to become largest economical plant in U.S. history, greatly outweighing the investments located to support another immense American infrastructure projects, specified as railways, highways and hydraulics for the internet.
Total investments in data centres and related infrastructure of artificial intelligence estimated at $10.3 trillion between 2025 and 2032, according to fresh estimates by economist Stijna van Nieuwerburgh published by Brookings Institution. This is simply a stunning 3.6% gross home product per year, on average. Never before has the US economy been so dependent on the improvement of 1 industry...
The flood of money spent on the expansion of data centres is simply a clear point in hard times for the construction industry.
By July of this year, seasonally adjusted $37 billion was spent on building private data centers — about 9 billion more than in the first 7 months of last year, according to the Department of Commerce.
Private spending on building everything else—houses, housing blocks, buying malls and so on—was about $46 billion lower than the year before in the first 7 months of this year...
Analysts estimation that capital expenditure of 5 alleged hyperscalers — Alphabet, Amazon.com, Meta Platforms, Microsoft and Oracle — They're going to be $4.2 trillion in 4 years, ending in 2029. A year, according to FactSet. The increasing part of these expenses is financed by debt.
Such spending increases the hazard to the financial sector if the boom breaks down. Van Nieuwerburgh said technology firms frequently usage off-balance sheet entities to lend from banks and private credit companies, and specified transactions are usually subject to tiny public reporting. This practice makes it hard to find how advanced the financial risks are, he added. If AI doesn't make adequate gross to pay off the debt created to build data centers, the effects could spread to the full financial system...
The growth driven by AI has led to massive increases in wealth on the stock market. According to the national Reserve, for the second quarter, US shares and investment funds were $63 trillion — Almost twice as much as at the end of 2022. This trend fueled consumer spending even erstwhile wage increases adjusted for inflation weakened. These gains were peculiarly visible to the affluent who tended to hold more of the net equity assets than the mediate class.
USA. Data center spending is larger than canals, railways and networks combined — creating jobs and wealth, but besides expanding inflation
Konrad Putzier and Justin Lahart • The Wall Street Journal • September 24, 2026 • 1,400 words
Last period another crucial article WSJ He reported that our biggest tech giants hide most of their massive investments in AI, holding billions of dollars of their financial liabilities in vehicles off balance sheet.
Every 4th large technology companies uncover their immense investment expenditure on artificial intelligence infrastructure, from data centres for integrated circuits.
But these numbers don't return the full scope of future expenditure, to which Google has committed, parent Alphabet, Meta Platforms, Oracle And many others. This is due to the fact that a immense part their upcoming commitments financial is not reflected in the balance sheets.
Nine leading technology companies had liabilities outside the balance sheet of about $3 trillion, mostly related to artificial intelligence, according to an analysis of footnotes in their latest securities reports published by the Wall Street Journal. These commitments grow faster than the conventional "capex" that over the past year, as reported, amounted to around $600 billion and were about 3 times larger than companies are blamed on late leases and long-term loans.
American technology companies with blue-chip place immense bets, based on assumptions about the request for AI calculations — and the availability of AI equipment — in a fewer years. They hope that they will easy meet all their obligations through future revenues erstwhile consumers and businesses implement AI in all aspect of America's life.
If these assumptions regarding technology and request prove to be incorrect, these agreements to safe future production capacity can become a immense burden for technology companies and their investors...
In all companies analysed by the Journal, the promises of payments under these uninitiated leases amounted to a full of $1.2 trillion of liabilities out of balance sheet, which is about 4 times more than the year before. In addition to Meta, the magazine has analyzed commitments to Alphabet, Amazon.com, Microsoft, Oracle, Nvidia, Broadcom, SpaceX and Advanced Micro Devices.
Data centers are filled multiple devices, including Nvidia systems utilized to train and run models and memory retention systems. To buy all this, companies sign long-term contracts well in advance to safe production with their suppliers.
These and another acquisition commitments in the companies that the diary investigated are as much as $1.9 trillion. According to accounting rules, acquisition obligations usually stay out of balance until the product or service is delivered...
The acquisition and contract commitments of Alphabet increased sharply and amounted to $811 billion on June 30. As with another companies, it is hard to tell precisely what it intends to buy on the basis of disclosures. The company has stated that these obligations relate mainly to "technical infrastructure and stocks" and "energy safety agreements for usage in data centres."
The Alphabet besides did not explain why these commitments rose so much compared to the $332 billion he had given 3 months earlier. The commitments cover respective years, and the obligations arising from the energy agreements apply even until 2054...
For those more afraid on Wall Street, the alarming signal is that any technology companies, which erstwhile seemed to have fortified balance sheets, had to frequently scope for capital markets.
Alphabet and Amazon late published results showing negative free cash flow, which means that their capital expenditures have exceeded the funds obtained from their activities.
And before we consider the implications of billions of commitments outside budget. Regardless of whether gross will always come, acquisition commitments and signed leases cannot be cancelled in the majority.
If something goes wrong, technology companies will pay advanced bills for infrastructure that they cannot usage economically. These commitments can besides make increasingly indebted companies gotta borrow even more.
"As these off-balance sheet commitments become increasingly frequent, larger and more complex, investors find it increasingly hard to measure the overall debt possible of companies," said analysts to book Morgan Stanley in April.
? costly commitments on the lease of data centres and chips are not seen in the balance sheets
Peter Rudegeair and Peter Santilli • The Wall Street Journal • August 16, 2026 • 1,100 words
From almost the earliest years Google has always been highly profitable, and this besides continued after changing its name to Alphabet, remaining 1 of the most profitable companies in the world.
But now, after nearly 30 years of existence, the vast sums spent on investments in the AI abruptly made it negative for the first time, which forced him to incur immense debts to cover his expenses.
Thanks to its super-voting actions, co-founder Larry Page controls the company together with Sergey Brin and reportedly tells people on Google that is "ready to go bankrupt than lose" the AI race.
So in a good way, America has put its full economy on the wonders of the AI revolution, and our biggest tech giants are doing something similar.
Over the past fewer months, I have watched interviews and read articles from many skeptics who have been convincingly arguing that they will most likely be lost bets.
A fewer months ago I discussed this situation in article, stressing work highly Fertile Criticism the financing of the AI industry, Ed Zitron.
In May, he stressed that a immense part of the gross backlog of our largest technology companies came from Anthropic and OpenAI, companies that are causing losses, which themselves powerfully finance:
The article The Information besides contained a fascinating illustration showing that about 50% of Amazon, Google and Microsoft backlogs (including all revenue, not only AI) — which is simply a stunning amount — is gross from OpenAI and Anthropic:
Just 2 weeks ago. both Amazon, and Google committed to invest a full of $65 billion in Anthropic, a company that In February, she raised $30 billion and plans to rise another 50 billionafter Amazon's investment in OpenAI worth $15 billion (or even $35 billion more) in February.
This is not what you do erstwhile there is simply a real, crucial request for AI services. Assuming these rounds will be closed at higher limits, this means Google has invested $43 billion and Amazon $33 billion in keeping Anthropic alive.
As I explained, most of the gross from AI from Google, Microsoft and Amazon comes from 2 companies that lose billions of dollars a year, have no way to profitability and can go on pay These companies only because, that companies (and investors) still They're getting the money.
These relationships are a completely toxic and deliberate effort to mislead investors and the general public.
As I explained then.
Therefore, the technological giants invested immense sums in OpenAI and Anthropic, which then return the same money in the form of payment for services. Similarly, most payments to Anthropic and OpenAI come from another AI startups having losses, whose venture capital investments represent income for these larger companies. quite a few this reminds me of the last stages of the Dotcom bubble.
giant breakdown Are you on AI technology stock?
Ron Unz • The Unz Review • June 1, 2026 • 8,700 words
Sometimes it's called "circular financing," and the 3400-word Wikipedia website about AI bubbles presents a good diagram illustrating 1 of these examples.
In this way, companies can artificially increase their own revenues, straight or indirectly subsidizing their customers. Moreover, if they take over the ownership of these customers, their profits increase as the valuation of the second increases.
The combination of artificially increased revenues and artificially increased profits gives a seemingly coherent image of good financial condition, but this is little than realistic.
None of the many critics claim that all this is illegal, but they claim that this kind of financial engineering can be very confusing to investors and origin massive bubble inflatation, while all these factors would abruptly turn around if the wind changed.
Interestingly, most of these financial critics tend to disregard claims that AI systems can exterminate humanity or that data centers must be destroyed to save our species. Instead, they claim that AI companies will most likely just go bankrupt, and their unfinished data centers can be converted into homeless shelters or simply devastated by copper wires.
Popular YouTube channel named The Tech Report He specializes in this kind of critical coverage of the AI manufacture and in fresh months I have watched quite a few their interviews in podcasts. Their guests are much more acquainted with these method and financial issues than I could always expect, and their films frequently gain hundreds of thousands of views.
Although most of these guys were people whose names meant nothing to me before, they interviewed a fewer days ago. with Dean Baker, a prominent economist who comfortly linked the current AI boom with what he saw and experienced during dotcom bubble and mortgage.
One of the remarkable facts that late came out is that 80% of Anthropic and OpenAI revenues came from just 1% of their customersWhich was a very susceptible situation. Baker noted that this 1% surely represents their most technologically advanced customers, and since so many leading AI models seem to get closer to their power, they could easy control to much cheaper AI systems, including the highly inexpensive Open-Weight AI models offered by Chinese companies, and late by Meta. If this were to happen, Anthropic and OpenAI could face serious financial problems, and the threat of specified a change would reduce their price power.
Another seemingly acquainted AI technologist, calling himself "Eli the Computer Guy", put out very akin arguments last week, arguing that OpenAI and Anthropic could easy lose 90% of their gross due to this. Both of these private companies inactive reportedly evidence multi-billion-year losses, declaring marketplace value between 1.5 and $2 trillion each, so any crucial drop in gross can be disastrous.
Ed Zitron is 1 of the most common and popular visitors of this channel, and in his latest podcast he argued that 1 of his most shocking claims had already been mostly verified.
For months he claimed that despite boasting and misleading press releases of all technological giants, the construction and commissioning of their fresh data centers was very late in time, and only a fraction of these systems were actually launched.
But all these companies have already reported that they have spent hundreds of billions of dollars buying ultra costly Nvidia AI chips that would fill these unfinished data centers. This led him to fishy that all these chips and racks with them most likely inactive lie in Taiwanese magazines, collecting dust, possibly already became obsolete before they were yet installed and activated.
Nvidia's worth over $5 trillionThe most valuable company in the world. But if in practice she sold many or most of her AI chips with pre-premier for one, 2 or 3 years, then her future gross sources must be much little safe than most people think. Meanwhile, the financial settlement of hundreds of billions of dollars of agreements that have not been installed and may never be utilized can seriously harm the balances of technology giants worth many trillions of dollars that have bought them.
According to popular belief, these advanced AI systems are desperately inaccessible as oil of our fresh industrial revolution. But what happens to the perception of the AI manufacture on the marketplace if Zitron is right and most of them are inactive unused in Taiwanese magazines for years?
As co-founder of Oracle, Larry Ellison is simply a billionaire for over 30 years, taking place among the richest people in the planet for most of that time, and in September 2025 very briefly reached position #1.
He was besides a major donor of the Israeli IDF, and late financed the creation of his boy David's vast media empire, including the acquisition of CBS, and besides bought control of TikTok.
However, most of Ellison's assets are based on his shares in Oracle, and for most of last year Zitron argued that there is simply a reasonable probability that Oracle would go bankrupt due to financial commitments to OpenAI. What's more, Ellison made about $20 billion of individual loans secured by his Oracle shares.
So if Oracle's stock became worthless, Ellison could go from the richest man in the planet in 2025 to a individual bankrupt in 2026, and his shares in TikTok, CBS and many another media would be auctioned during the sale, along with his private island in Hawaii and another things.
In the last fewer days Ellison committed to an additional $9.2 billion share Oracle as collateral for the loanat the same time Suddenly canceling plans to sale these shares at $7.5 billion.
Perhaps accidentally, according to the latest intelligence for Zitron, financial ice begins to crack underground Oracle.
https://www.unz.com/runz/america-pravda-will-ai-exterminate-the-human-species/
