finance on Dark Edge

The AI bubble is going to pop

darkedge@darkedge.world (Guido Stevens) — Thu, 07 Aug 2025 00:00:00 +0200

A summary of the links shared in a fascinating Mastodon thread about the current AI hype cycle and the bubble that’s about to pop.

What’ll Happen If We Spend Nearly $3tn on Data Centres No One Needs?

The FT features a ferocious takedown (Elder 2025) (paywalled, open archive version here), of a Morgan Stanley AI booster briefing. It opens with a 1990s-themed web 1.0 graphic featuring a portrait of Sam Altman to set the tone.

FT quotes a wildly optimistic Morgan Stanley briefing:

GenAI revenues could exceed $1tr by 2028, with close to 70% variable margins, compared to just $45bn in 2024. […] ROI of AI should already be positive this year, generating $50bn in revenues, and that this will grow to exceed $1tr/year by 2028

To put that into perspective, that’s in the ballpark of half of all S&P500 investment!

Hyperscaler funding of $300bn to $400bn a year compares with annual capex last year for all S&P 500 companies of about $950bn.

After the telecoms bubble, the money was gone. But at least we still had all the internet infrastructure built with that money. This time may be different. Nvidia GPUs are not general-purpose infrastructure that are useful for any type of compute; they’re specifically optimized for running Large Language Models. And they obsolete really fast.

Morgan Stanley estimates that $1.3tn of data centre capex will pay for land, buildings and fit-out expenses. The remaining $1.6tn is to buy GPUs from Nvidia and others. Smarter people than us can work out how to securitise an asset that loses 30 per cent of its value every year, and good luck to them.

Even worse: those GPUs need a firehose of electricity to run. Even if the current owners go bankrupt, anybody who wants to pick up those assets is going to be faced with the same operational problem: to pay for the electricity. That is, assuming you’ll be able to buy it. Which is a wild assumption in itself. And it’s where the AI hype runs into the unfolding climate catastrophe really hard.

America needs to find an extra 45GW for its data farms, says Morgan Stanley. That’s equivalent to about 10 per cent of all current US generation capacity, or “23 Hoover Dams”, it says.

Build out power capacity to the tune of 23 Hoover Dams in the coming 3 years? Never gonna happen.

Back to the investment. How is all that money supposed to be earned back? It’s worth zooming in on risk embedded in the assumption made by Morgan Stanley, of outrageous revenue growth, from $45bn in 2024 to $1tn in 2028.

When the base case is for 1,900 per cent revenue growth by 2028, isn’t it worth considering the risk of a shortfall?

The article compares the current investment cycle in AI with the telecoms bubble of the late 1990s. Except, this is much bigger. As in: 10 times bigger.

In 2000, at the telecoms bubble’s peak, communications equipment spending topped out at $135bn annualised. The internet hasn’t disappeared, but most of the money did. […] Peak data centre spend this time around might be 10 times higher

Sounds like a repeat of the GFC waiting to happen, to me.

The AI Bubble Is so Big It’s Propping up the US Economy (for Now)

To put that into context: Brian Merchant points out that this investment level is so high, that it’s the main prop keeping the US economy afloat. (Merchant 2025)

He analyses a Wall Street Journal report that makes his head spin.

I’ll just repeat that. Over the last six months, capital expenditures on AI—counting just information processing equipment and software, by the way—added more to the growth of the US economy than all consumer spending combined. You can just pull any of those quotes out—spending on IT for AI is so big it might be making up for economic losses from the tariffs, serving as a private sector stimulus program.

The Hater’s Guide To The AI Bubble

Ed Zitron has been banging this drum for a while. He makes additional points. So much, in fact, that you should read his whole article (Zitron 2025).

He does a drilldown on the financials, and concludes the revenue is puny and partly double-counted. There is no profit.

Yes, generative AI has functionality. There are coding products and search products that people like and pay for. As I have discussed above, none of these companies are profitable, and until one of them is profitable, generative AI-based companies are not real businesses.

I believe that the generative AI market is a $50 billion revenue industry masquerading as a $1 trillion one, and the media is helping.

Zitron concludes, that the stockmarket depends on Nvidia keeping up its rocket growth rate. Which in turn depends on the hyperscalers keeping up their investment. While none of them actually make money on AI.

We are in a bubble. Generative AI does not do the things that it’s being sold as doing, and the things it can actually do aren’t the kind of things that create business returns, automate labor, or really do much more than one extension of a cloud software platform. The money isn’t there, the users aren’t there, every company seems to lose money and some companies lose so much money that it’s impossible to tell how they’ll survive.

References

Elder, Bryce. 2025. “What’ll Happen If We Spend Nearly \$3tn on Data Centres No One Needs?” Financial Times, July. https://www.ft.com/content/7052c560-4f31-4f45-bed0-cbc84453b3ce.

Merchant, Brian. 2025. “The AI Bubble Is so Big It’s Propping up the US Economy (for Now).” https://www.bloodinthemachine.com/p/the-ai-bubble-is-so-big-its-propping.

Zitron, Edward. 2025. “The Hater’s Guide To The AI Bubble.” Ed Zitron’s Where’s Your Ed at. https://www.wheresyoured.at/the-haters-gui/.

Posthypercapitalism

darkedge@darkedge.world (Guido Stevens) — Sun, 05 Oct 2008 00:00:00 +0200

The Great Financial Crisis of 2008 makes clear that in a networked world, we need new, networked ways of thinking about the world.

Figure 1: New York Stock Exchange

This is an old article, restored from a previous blog.

Rorschach effect

The crisis on Wall Street is like a Rorschach test: it seduces people into making statements that primarily reflect their own state of mind. Everybody finds something to his liking that he latches onto.

Germans, with characteristic bombast, are talking about an economic 9/11 self-inflicted on Wall Street by investment bankers turned suicide attackers.

The French, in typical sweeping language, are proclaiming the need to re-invent capitalism. Sarko will unveil proposals for a new world order on a summit later this year.

Chinese media, without apparent sense of irony, point to excess creation of US liquidity and debt. Which are hugely financed by Chinese government dollar holdings.

American conservatives decry massive government intervention as ‘socialist’ and ‘un-American’. In other news, 25 billion dollars of taxpayer money has just been awarded to the US car industry for failing to anticipate fuel-efficiency demand.

Closet socialists, populists and just about everybody is tumbling over each other to condemn Wall Street greed and fat cat bonuses.

Economists, meanwhile, are furiously debating what the heck is going on. Is this a liquidity crisis, or a solvency crisis? Isn’t it insane to throw a trillion dollars at a problem we don’t understand? Isn’t it irresponsible not to?

The problem is: everybody is right – well, except maybe the Germans :-). The causes of this crisis are myriad: greed and corruption, faulty decisions, faulty technical models, excess liquidity, liquidity contraction, insolvency, intransparancy, lack of oversight and regulation, absurd levels of leverage, you name it.

If everybody is right, that probably means nobody really understands what is going on. There’s a different perspective that doesn’t get wrapped up in the inner workings of the financial system.

nonlinear complexity

Surprisingly effective explanations for the current financial meltdown emerge from the natural sciences. A growing body of work translates insights from biology, physics and mathematics into powerful models for economic interactions.

The key element that binds these narratives together is: emergent complexity. These models skimp on explaining individual behavior. They lack a ’theory of the firm’ and ‘bounded rationality’ concepts. Indeed, such models feature the coarsest imaginable agents, with only very rude binary (positive/negative) relationships to other agents and very rude binary (alive/dead) state.

Stringing such simplex agents together in networks that obey equally simple rules, modeling outcomes are achieved that show an uncanny resemblance to actual, historical, economic data time series. The implications are profound: individual decisions don’t matter very much, the actual outcomes are determined by structural properties, i.e. by the network of interactions.

Even more mind-boggling is the cross-disciplinary reach of these effects: a stock market crash very much resembles a traffic jam very much resembles species extinction events: the mathematics is much the same in each of these very different problem domains.

This points to an underlying regularity in the laws governing complex systems, of which the economic system is but a specific manifestation. To paraphrase McLuhan: the network is the effect. It is the structure of a network, rather than the actions of network participants, that determines the eventual outcome.

fractal markets

Benoit Mandelbrot, the mathematician of fractal fame, performed a technical (mathematical) analysis of market behavior in general, and the 1998 collapse of the Long Term Capital Management hedge fund in particular. The Misbehavior of Markets: A Fractal View of Risk, Ruin and Reward was published in 2004. (Mandelbrot and Hudson 2004) His conclusion? Markets don’t behave “normally”. In statistical terms, market volatility does not conform to a “normal” Bell curve. Extreme market events occur much more frequently. The under-estimation of the likelihood of extreme market events was a fatal modeling flaw that broke LTCM in 1998. And, we might add, is breaking banks all over the place right now.

creative extinction

Paul Ormerod, an economist, turns to biology to explain Why Most Things Fail. (Ormerod 2006) He describes patterns of failure and the similarities between biological failures (species extinctions) and economic failures (bankruptcies). Like Mandelbrot, he is keenly aware that extinction events are not governed by a “normal” distribution but by a “power law” distribution, which results in much higher probabilities for high-impact disruptions (mass extinctions, economic crashes).

A key insight to emerge from his analysis, is the extremely limited amount of information agents really have about their environments. The mathematical model fits observed reality only if agents have near zero “cognitive ability”. Translation: the sorry state of the world can only really be explained by our stupidity :-).

In short, despite the ability of humans and human institutions to act with intent, in reality it is as if they operate close to the paradigm of the agent with zero cognitive ability.

[…]

The clear implication of this abstract theoretical model is that agents, firms, individuals, governments have very limited capacities to acquire knowledge about the true impact either of their strategies on others or of others on them.

(Ormerod 2006)

Doesn’t that sound awfully much like an adequate description of the mess the CDO and CDS trades have created? Banks have stumbled into a situation in which they’re unable to adequately judge their own or their peers’ risk exposure.

flocks of black swans

Scores of other books in the non-fiction section of your local bookstore elaborate the same meme: complexity gives rise to catastrophe. A world of “normal accidents” is much more unsafe than we’d all like it to be.

There’s a bit of hope, however, in the realization that stock market crashes and Katrina-category extreme weather events are manifestations of similar patterns of interdependencies and feedback loops.

Wall Street crashes and must be rescued by huge public policy interventions. This powerfully underscores the limits of pure market-driven coordination, even in coordinating the sanity of markets themselves. Maybe this forceful demolishing of entrenched ideological doctrines creates an opportunity for some fresh insights to emerge, and to be applied to the wide range of problems, economical and ecological, that we’re facing this century.

References

Mandelbrot, Benoit B., and Richard L. Hudson. 2004. The (Mis)Behavior of Markets. New York: Basic Books.

Ormerod, Paul. 2006. Why Most Things Fail. London: Faber & Faber.