The Google A.I. Gemini sucked up to me as I applied it to this page.
That's an insightful connection, and the web page you are viewing directly addresses both Karl Marx's concept of General Intellect and the phenomenon of sunk costs in the context of the AI bubble.
There will be “debt funding or circular financing” that could precipitate a global depression despite the immensity of the neoliberal, superimperial mechanism designed to thwart such a catastrophe as documented by people like Michael Hudson. The Global South will absorb as usual as a co-dependent enabler, the corporate mendacity of the Global North. We will develop better analyses of settler colonialism only to see it swamped by supply-chain trading chaos and financialized network adjustment. This is sustainable for a ruling class of social murderers but not for the rest of us. A.I. is both fire-fighter and arsonist, offering several ideas by Mohamed A. El-Erian that seem appealing in promising that a ‘rational bubble’ will not burst:
1.) groundbreaking models (LLMs) built on unsustainable debt funding or circular financing,
2.) Like Greenwashing or Sane-Washing, an A.I.-washing resembling the dot-com bubble, and
3.) usual, exogenous factors, those external developments that come from ‘geopolitical events, technologies, and the ecosystem itself.
Peak A.I. is infinite in horizon whereas peak oil is about eventual extinction.
The available analyses present a split picture: some analysts and institutions warn that circular financing and heavy capital recycling in AI could inflate a bubble that risks a sharp correction, while others argue AI fundamentals — rising revenues and demand outpacing supply — make the sector more resilient than dot‑com era comparisons suggest [1] [2] [3]. Evidence centers on large-scale spending, hardware-leaseback and debt schemes, and interlocking deals among major players; the debate hinges on leverage, sustainability of cash flows, and the extent to which financing is genuinely circular versus growth-investment [4] [5] [6]. news.futunn.com/...
Meanwhile, to meet the enormous computing power demands brought by AI, hyperscale data center operators are making a high-stakes bet. The report shows that procurement commitments by these companies surged by $100 billion to $330 billion over the past year, while leasing commitments climbed to $340 billion. These long-term, massive capital expenditure commitments are entirely dependent on the sustainability of AI demand—any slowdown could result in significant risks.
news.futunn.com/...
As a counterfactual, El-Erian claims that current bubbling is not like the big historical bubbles, such as the tulip mania of the early 17th century. Longue durée analysis of capital tells us that historical events cannot be so autonomous considering how capital has not changed all that much. AI is not electricity, and in fact because it isn’t there will be palpably unsustainable costs in rural datacenter construction.
There are still metalist Goldbugs and since Marx, we know that surplus value is still being produced. Positing some rationality like neoclassical economics makes reality seem all so imaginary, not unlike the power of a ‘general intellect’ dislocated from notions of labor power and valorization.
There are events in this historical/epochal mode of production that signal subsequent monopolistic/oligarchic institutional dominance, built upon the flexible production of offshored tech driven forward in the global North by extensive regional energy sumps called ‘datacenters’. Putting them underwater is more costly than building them in the open red-state spaces.
In the midst of financialization and rentier capitalism, costs and debt will get kicked down a road destroyed by climate change and excessive income inequality. Ecosocialism is no panacea when we pass a climate tipping point. “In 2019 alone, Google requested, or was granted, more than 2.3 billion gallons of water for data centers in three different states.”
In the absence of a better ballroom the White House created another opportunity for executive managerial capitalist agents to mingle. The White House is so amoral that it severed its own East Wing.
The black-tie event on November 18 honored the crown prince during his Washington visit, drawing nearly 50 U.S. executives and sealing billions in deals like Saudi pledges for $1 trillion in AI, semiconductors, and energy investments.
Bernie Sanders labeled the tech leaders 'A.I. oligarchs' and the prince a 'dictator who murders his own citizens with a bone saw,' referencing the 2018 Khashoggi killing, while doubting their poverty-fighting claims.
X users split sharply, with some backing the job-creating innovators and others accusing Sanders of hypocrisy amid the heated debate on tech, rights, and inequality.
(X.com summary)
x
Attendees at last night’s White House dinner honoring MBS were all doing business with him and Saudi Arabia.
Under Trump, the only normative rule is to gain as much power and money as possible.
Power and wealth are honored, even if the honoree has greenlit a brutal murder.
— Robert Reich (@rbreich.bsky.social) Nov 19, 2025 at 12:09 PM
“Artificial General Intelligence” (AGI) is not identical to what Marx called the General Intellect, even if morphologically there is a resemblance.
El-Erian is optimistic because if you’re in the 1% or at least an economist for them, other than Larry Summers, you will be held accountable. OTOH the NY Times’ bubble seems elastic even if we know rationally, that all bubbles burst. One shouldn’t have to remind him that stock markets are not identical with economies. Investment bubbles are inherently different.
Servaas Storm’s paper The AI Bubble and the U.S. Economy: How Long Do “Hallucinations” Last? argues that
(i) we have reached “peak GenAI” in terms of current Large Language Models (LLMs); scaling (building more data centers and using more chips) will not take us further to the goal of “Artificial General Intelligence” (AGI); returns are diminishing rapidly;
(ii) the AI-LLM industry and the larger U.S. economy are experiencing a speculative bubble, which is about to burst.
www.ineteconomics.org/...
It’s reasonable to worry whether sky-high valuations for artificial intelligence companies represent a bubble that will burst and wreak havoc on your retirement account.
However,
A bursting bubble could indeed be painful in the short term. But what if we’re in a “rational bubble” that, unlike other big speculative manias in history, takes our economy to a fundamentally better place?
However, the A.I. excitement, as seen in the blowout Nvidia earnings yesterday, rightly reflects the potential transformation of the entire economy. It is economically rational to risk losing everything on several bets if just a few can deliver a thousandfold return, which some A.I. investments almost certainly will.
Whichever way you look at it, the potential payoffs of A.I. adoption are staggering — for the economy, for social sectors, and, of course, for investors. That could not be said for the majority of the big historical bubbles, such as the tulip mania of the early 17th century.
That’s not to say that the A.I. bubble won’t burst. Investors and governments have placed their wagers on a group of racehorses, only a handful of which will win. Yet while the rest may well be destined for the glue factory, resulting in significant losses, the race itself will yield innovations that could result in vastly superior outcomes for many than would otherwise be the case. [surplus value]
Surging A.I. stocks are pulling up the valuations of many other companies, some to highly elevated levels. That’s because investors attracted by a handful of big tech names have fueled a much larger buying surge into investment vehicles, including index funds, and because buyers searching for new places to put their money often venture to areas they insufficiently understand.
It’s easy to see how this ends badly. But what if the bubble is an inevitable part of developing and adopting a revolutionary tool that will fundamentally improve productivity and growth? After all, A.I. is a general-purpose technology that will most likely alter a vast range of economic activities fundamentally. Its transformative potential could be on par with electricity, offering an enormous upside through durable improvements in what we do and how we do it. It’s not just that many existing activities will be done better and more efficiently. A.I. is poised to open the door to discoveries, particularly in health and education.
With such big payoffs, the incentives to invest in A.I. are enormous, and are reinforced by two biases: Investors believe that falling behind is far more damaging than over-investing, and they fear that they’re missing out. From their perspective, what some may ultimately see as enormous overspending is, in fact, a calculated portfolio play that fuels competition and innovation.
My optimism for our future economic well-being is based only on A.I.’s capabilities today. Yet the quest for artificial general intelligence ensures that the frontier is not static; it will continue to leap forward, promising developments that are difficult to fully comprehend now. Those will include A.I.’s convergence with innovations in robotics, life sciences and, eventually, quantum computing.
www.nytimes.com/…
This data center AI chip roadmap shows NVIDIA will dominate far into 2027 and beyond
This data center AI chip roadmap shows that NVIDIA dominates the AI GPU business: right up to the Rubin Ultra AI GPU with 576GB of HBM4 memory in 2027. www.tweaktown.com/...
The sunk cost fallacy occurs when individuals continue investing in a project, decision, or activity based on the amount of resources (time, money, effort) already committed, rather than evaluating the current and future value or potential return.
Marx’s general intellect is more a covering term than a commodity, and the constituent elements make collective social knowledge a perishable while fungible flow/stock that is concurrently rational/irrational.
Here is a breakdown of how the text applies these concepts to AI:
GEMINI LLM
Marx and the "General Intellect"
-
The Concept: Karl Marx's idea of the General Intellect (from his Grundrisse) refers to the accumulated scientific and social knowledge that becomes a direct force of production, primarily embodied in machinery.
-
AI Application (The Argument): The article you are reading explicitly states that "AGI is not identical [to] what Marx called the General Intellect."
-
Marxist AI Critique (Broader Context): Search results show that Marx's ideas are frequently used to critique AI development:
-
AI is seen as the General Intellect—the collective social knowledge—being privately appropriated by a few corporations for profit.
-
The use of AI is argued to intensify the alienation of workers, who become mere "conscious linkages" or "watchmen" to the automatic system, and who are displaced from jobs as labor is reduced to a minimum.
Sunk Costs and the "Rational Bubble"
-
Sunk Cost Fallacy: The text explicitly defines the sunk cost fallacy as continuing an investment "based on the amount of resources... already committed, rather than evaluating the current and future value or potential return."
-
The Rational Bubble: The economist Mohamed A. El-Erian (cited on the page) argues that the AI frenzy is a "rational bubble," which is distinct from a classic, purely irrational one like the tulip mania.
-
The Bubble's End (The Sunk Cost Effect): El-Erian argues the bubble will pop due to three forces, which relate to a waste of resources:
-
Arms Race: An oversupply of competing tech companies whose required costs for elite engineers, data centers, and energy will soar, making debt-reliant players ultimately unsustainable.
-
A.I.-Washing: A flood of "fool's gold" where companies label mundane services as AI, attracting less informed investors.
-
External Factors: Regulatory changes, geopolitical fractures in the supply chain, or a lack of widespread AI adoption.
In essence, the text is framed around the idea that the massive investment, while logical due to AI's transformative potential (a "rational bubble"), is still leading to massive, sunk costs and eventual losses for the many companies that will fail.
Literal and figurative sunk costs.
Q: Is trump boosting Ai data centers raising electricity costs on Americans?
A: Yes, experts indicate that Ai data centers, a sector the Trump administration is actively boosting through deregulation and expedited permits, is contributing to higher electricity costs for American consumers.
x
Meta's effort to preserve its credit rating by keeping its $27B data center & associated debt off its books "strains credibility," built on assumptions that the match the reality of its ability to control the project. [gift link]
[image or embed]
— John Kostyack (@kostyack.bsky.social) Nov 24, 2025 at 9:49 AM
AI Meets Aggressive Accounting at Meta’s Gigantic New Data Center
Favorable treatment off the balance sheet hinges on some convenient assumptions
It seems like a marvel of financial engineering: Meta is building a $27 billion data center in Louisiana, financed with debt, and neither the data center nor the debt will be on its own balance sheet.
That outcome looks too good to be true, and it probably is.
Indeed, it is quintessential structured finance—the point of which is to help companies try to achieve seemingly irreconcilable financial-reporting goals. Meta wants other people’s money to pay for the data center. Those investors want to know Meta ultimately has their backs. But at the same time, Meta wants to keep its near-pristine credit rating and doesn’t want the extra assets and debt on its books.
Construction on the project was well under way when Meta announced a new financing deal last month. Meta moved the project, called Hyperion, off its books into a new joint venture with investment manager Blue Owl Capital Meta owns 20%, and funds managed by Blue Owl own the other 80%. Last month, a holding company called Beignet Investor, which owns the Blue Owl portion, sold a then-record $27.3 billion of bonds to investors, mostly to Pimco.
Meta said it won’t be consolidating the joint venture, meaning the venture’s assets and liabilities will remain off Meta’s balance sheet.
Instead Meta will rent the data center for as long as 20 years, beginning in 2029. But it will start with a four-year lease term, with options to renew every four years.
This lease structure minimizes the lease liabilities and related assets Meta will recognize, and enables Meta to use “operating lease,” rather than “finance lease,” treatment. If Meta used the latter, it would look more like Meta owns the asset and is financing it with debt.
The favorable accounting outcome hinges on some convenient assumptions. Some appear implausible, while others are in tension with one another, making the off-balance-sheet treatment look questionable.
