In terms of concentration risk, the AI trade’s now on par with the largest bubbles of the past 150 years.
I mentioned this a few days ago, but I’ll reiterate it here: The “Big 10” AI names, defined as the Mag7 plus AMD, Broadcom and Micron, now comprise more than 40% of overall US equity market cap.
That’s comparable to, or in some instances slightly larger than, the share of market cap commanded by the leadership during four previous bubbles. The exception’s the railroad boom in the latter half of the 19th century, when related stocks at one point comprised nearly two-thirds of market cap in America.
The (mordant) joke writes itself: That’s a lot of upside! For the AI names, I mean. Ostensibly, they can claim another ~20ppt worth of market cap share between them before the train flies off the tracks.
Snark aside, I wouldn’t rule such a scenario out. If Micron’s sales guide is any indication, demand for crucial AI components remains insatiable. Bloomberg reported Friday that Broadcom’s lining up $60 billion in new financing to help Anthropic access compute as the company gears up for a historic public offering before Thanksgiving. And on and on.
Assuming there’s more gas in the tank (coal in the tender) for the AI trade and that the associated concentration risk rises further towards the railroad boom extremes, it’s worth exploring the analogue.
The railroad bubble, BofA’s Michael Hartnett reminded investors in his latest weekly, came in two waves. One lasted from 1861 to 1872, the next from 1877 to 1881. At the peak during the first wave, spending on the buildout reached 5% of US GDP, before European financing dried up ahead of the 1873 bust.
If you assume hyper-scaler capex of between $1.2 trillion and $1.4 trillion in 2027, that’d be less than 5% of contemporaneous US GDP, which is to say short of the railroad boom peak (and, Hartnett noted, lower than the same metric for the fiber-optic network buildout in late-1990s).
So that’s — I’m not sure what the right word is — “comforting,” I guess? More importantly, prices for “bubble inputs,” if you like, are still rising, and rather dramatically at that. Just ask HBM margins.
Soaring prices are in stark contrast to the conditions which prevailed at the end of the second phase of the railroad boom when, as Hartnett recounted, excess capacity and freight-rate deflation meant “revenues and profits were unable to sustain the capex boom.”
The bad news, which is to say the argument for why the AI bubble may be living on borrowed time and thereby won’t likely match the railroad boom on sundry metrics of “revolution”ary excess, is straightforward: Rates are rising.
“The railroad booms were backstopped by an era of falling UST yields,” Hartnett wrote. “That’s evidently not the case today.” No, “evidently” not, and the irony is that the tsunami of IG debt supply to fund AI capex is a factor in explaining why long-end Treasury yields are perched at the highest levels in decades.
The figure above, from the same Hartnett note, shows you 10-year US Treasury yields alongside railroad bond yields for the period encompassing the two booms.
To reiterate what’s quickly becoming an ad nauseam refrain in these pages amid the staggering run-up in G7 bond yields: Rates will matter eventually.
For now, though, that seems a quaint notion in the face of a technology with the potential to save our species (as long as it doesn’t “decide” to kill us first in what, if you ask me, would be a welcome coup de grâce). As Hartnett quipped, describing the zeitgeist, “Rates don’t matter when you’re curing cancer.”
Rates should drop precipitously when Skynet becomes self-aware and sends the first Terminator…
wouldn’t freight rates more comparable to token cost than producer prices?
Rails got cooked in the first pop in UK because they raised rates, and over leverage on bets, so seems fair to look at that for US AI.
So far not a lot of concrete and detailed explanations of AI bust (kills us) or boom. Investing this is like navigating an unfamiliar port at night in the rain with no radar or GPS. On this occasion I see H like the guy you send up to the bow to yell when he sees something. Glad someone is awake and on watch.
Gotta have someone casting the lead line too!
I like the analogy, and I’ll add that we’re trying to pick out the channel markers, but it’s Christmas time and all the lights on shore are green and red.
The Railroad Boom ended spectacularly in the Panic of 1893. That led to a four-year depression which saw unemployment reach 25%. It was the worst depression the country had ever seen at that point, and it essentially set about the end of the Gilded Age. The panic began when the Philadelphia & Reading Railroad filed for bankruptcy in February of 1893 (yes, _that _ Reading Railroad), but the main trigger was actually the collapse of the National Cordage Company which held a monopoly in the rope and twine industry. They had borrowed to the hilt in order to try and corner the world hemp market and failed. Rumor spread that the company was out of cash, and was desperately seeking a $50,000 loan. Fearing losses, the banks began to call-in their loans in earnest, and National Cordage went bankrupt on May 4th, 1893. The next day, Friday, May 5th, (the original “Black Friday”) the stock market collapsed. The panic spread, as the banks then frantically tried to call-in loans to overextended railroads and some other industrial giants. Most were buried in debt and could not hope to cover their loans, and the ensuing wave of bankruptcies and bank closures ran through the summer of 1893.
Because there was no federal deposit insurance (FDIC) or a central bank, over 500 banks and 15,000 businesses closed permanently, wiping out the savings of millions of Americans. Monopolies, corruption, and a lack of banking regulations were later blamed, and led to a new age of muckraking, monopoly busting, regulation, and progressive reform (“Maybe we should get us some of that reform.”) The similarities are a little frightening. Let’s sincerely hope that banking regulations, the FED, and the FDIC mean that this particular history simply cannot repeat itself now.
Now we have government backstops for too big to fail, and especially considering the role of AI in defense/war.
“There is a massive global rush to buy .si domain names because the United States government decided to use “Super Intelligence” (SI) instead of “Artificial Intelligence” (AI).
What is a .si Domain?
• It is the official internet code for the country of Slovenia.
• ARNES, a network group in Slovenia, manages these web addresses.
• Anyone in the world can buy a .si web address.
Why the Sudden Rush?
• US President Donald Trump gave a speech at the United Nations on September 22, 2026.
• He said the US government would call artificial intelligence “Super Intelligence” or “SI”.
• He signed an official order on September 29, 2026, to make this change permanent for federal agencies.
• Because “SI” matches Slovenia’s internet code, people rushed to register matching web addresses.
The Numbers and Risks
• Slovenia only had about 3,500 new sign-ups in August 2026.
• In September 2026, people registered more than 46,000 new .si addresses.
• Total registrations passed 260,000.
• Most buyers are investors trying to resell the names for profit rather than building real websites.
• Experts warn that these similar letters might lead to phishing or typo mistakes next to .ai websites.
Whose trophy wife is from Slovenia?
This is such an informative post. Thank you.
You’re welcome. I recently saw a documentary on the end of the Gilded Age. At the time of the Panic of 1893, there were a number of industrial millionaires building spectacular mansions on 5th Avenue in New York. Exquisite designs, marble interiors, imported craftsmen, et cetera. When the market collapsed, nearly all of the work was halted immediately, and most of the mansions were never completed, and were eventually sold-off for other purposes (at a mere fraction of what they cost to build.) I know most people feel that kind of collapse simply could not happen again, but it is still worth understanding the mechanics of the thing, and the impact that followed.
I don’t know whether there will be an AI bust or not, but I’ve got to remark on how much kindling there is lying around with more arriving daily. Is that gasoline I smell. If somebody lights that leaning tower of AI debt, it could get intense.
The amount of FUD here is telling me there’s a lot of room to run. The Mag7, the most profitable companies the world have ever known seem far from insolvency, and AI is showing massive productivity gains. I’m more worried about the job prospects of recent college graduates, the plunge in births and the unraveling of society. But to each their own.
“most profitable companies the world have ever known seem far from insolvency”
at least until the write-downs start. Let me see, how many times has this happened to me.