Oct 16, 2026 · 59 min
AI’s data-center boom echoes the railroad crash of 1873
Will The AI Boom End Like The Railroad Crash Of 1873? — ft. Liaquat Ahamed
The episode tests whether debt-fueled AI infrastructure spending could produce the same overbuilding, contagion, and distressed-asset opportunities seen after 1873.
- 1Transformative technologies attract individually rational investments that can collectively create dangerous overcapacity.
- 2Debt dependence, crowding out, rising rates, and delayed data-center IPOs may signal a more fragile AI cycle.
- 3Historical parallels clarify systemic risks but cannot predict when a correction will arrive or which company will fail.
Don't miss
Liaquat Ahamed identifies the current market as manic while warning that earnings and real technological progress can allow excessive expectations to continue for years.
The brief
Financial historian Liaquat Ahamed and Ed Elson compare today’s data-center expansion with the railroad boom that preceded the Panic of 1873.
The parallel is not simply technological optimism: many cautious companies can collectively overbuild, increasingly relying on debt once internal resources run short.
Ahamed points to crowding out, rising interest rates, delayed data-center IPOs, and political resistance as signs that the AI infrastructure cycle is becoming more fragile.
History offers no reliable timetable, but the episode argues that crises often build slowly before arriving suddenly—and that financially strong companies can benefit afterward.
The standout tension is whether genuine AI progress can sustain manic valuations long enough to obscure the financial strain beneath the buildout.
What was said on this episode
30 statements · 5 positive · 19 negative · 3 mixed · 3 neutral
Global GDP doubled and trade increased fivefold during the 1850–1870 boom.
“GDP doubled in those 20 years. Trade went up five times.”
Listen at 3:35
Three-quarters of annual railroad investment was financed with borrowed money.
“So the $3 billion a year that eventually went into railroads, three quarters of that was borrowed money.”
Listen at 4:32
Within a year of Jay Cook’s failure, one-quarter of U.S. railroad companies defaulted.
“Within a year... A quarter of the railroad companies in the U.S. had defaulted, and it had this massive cascading effect across the market.”
Listen at 13:45
Germany’s move to gold and U.S. silver legislation caused silver prices to fall 30–40%.
“The combination of those two things leads silver prices to go down by 30-40%.”
Listen at 15:57
Monetary tightening transformed a normal downturn into a global depression and drove prices down 30–40%.
“And that is what transformed what was a normal down cycle into a global depression, which caused prices to fall by 30 to 40 percent.”
Listen at 16:48
The key risk in the data-center boom is debt-financed overbuilding.
“The key risk in the data center boom is overbuilding and overbuilding finance by debt.”
Listen at 20:51
NVIDIA is effectively lending to its customers to support data-center construction.
“At the center is a company like NVIDIA, which is essentially lending to its customers.”
Listen at 21:43
Data-center overbuilding may cause token prices to decline.
“And I suppose the analogy today is, you know, is... is all of this data center construction leading to a fall in the prices of their output, which I suppose is tokens, and token prices are going down.”
Listen at 22:44
Individually cautious technology companies collectively produce overbuilding.
“It's the collectivity of their decisions that leads to overbuilding”
Listen at 25:20
Competitive technology markets probably cannot prevent collective overbuilding.
“Probably not. I mean, I think it's just inherent in the competitive situation that everyone wants to be one of the two or three companies that comes out on top.”
Listen at 26:06
Historical private-investment booms lasted 2–4 years, created excess capacity, raised rates, and ended in crashes.
“All of them led to a jump in private investment of 2% to 4% of GDP. All of them lasted two to four years. All of them ended up with excess capacity. And all of them... interestingly, ended up with a spike in interest rates as everyone was competing for capital in the final stages of the boom. And all of them ended up with a crash.”
Listen at 27:14
The data-center boom may be halfway through a potentially four-year cycle.
“The data center boom has been going, let's say, for two years. We've had booms go on for four years. So we could be halfway through the process.”
Listen at 28:48
Projected next-year data-center investment is roughly 3–4% of GDP, below historical extremes.
“This current boom is not at 5% of GDP. Next year, we're projected to put in, I don't know, $1.5 trillion. That's probably 4% of GDP, 3% to 4%. We're not at the sort of extremes that we've experienced in the past.”
Listen at 29:27
Financial systems have limited capacity to finance more than 3–4% of GDP in private investment at once.
“It also causes, there's a sort of limited capacity to finance more than 3% to 4% of GDP. in one go.”
Listen at 30:35
Long-term U.S. interest rates have risen about 150 basis points amid the investment boom.
“Interest rates have risen, whatever, 150 basis points in the long end in the U.S., even more in other countries.”
Listen at 30:55
Eroding profitability is a major risk for data-center investments.
“The other big factor is the erosion of profitability.”
Listen at 33:38
Financial crises typically develop longer than expected and then unfold rapidly.
“these things go on for longer than you imagine, and once they happen, they'll happen much quicker than you thought they could.”
Listen at 36:11
The Rothschilds avoided speculative railroad fads and benefited from doing so.
“In Europe, the Rothschilds were brilliant. They had become rich by avoiding fads. So they avoided any of this.”
Listen at 36:28
The equity market currently equals 250% of GDP.
“the equity market is now 250% of GDP.”
Listen at 42:00
Recent economic growth is partly driven by wealth effects from rising equity prices.
“The economic boom that has sort of become self-sustaining in the last two years is in part fueled by wealth effects from the rise in equity prices.”
Listen at 42:05
A major equity-market decline could reduce consumption through negative wealth effects.
“If the equity market takes a tumble, what will be the wealth effects of equities going down on consumption?”
Listen at 42:34
A 6%-of-GDP budget deficit leaves limited fiscal-policy capacity.
“We've got a budget deficit that's at 6% of GDP. You know, we've probably run out of bullets in terms of fiscal policy.”
Listen at 42:55
Foreign investors put approximately $1 trillion into U.S. equities over the past year.
“a trillion dollars has gone from foreigners into the US equity market in the last 12 months.”
Listen at 43:14
The 1873 railroad crisis is a strong analogy for today’s AI boom.
“I think 1873 is a pretty good analogy.”
Listen at 46:04
Modern policymakers are not gold-standard constrained but may face inflation constraints.
“This time around, no one's tied to the gold standard. But are we tied to keeping inflation under control? Does that make the knock-on effects much worse?”
Listen at 46:44
Crazy magical thinking currently permeates financial markets.
“I think we've got plenty of examples of sort of crazy magical thinking that permeates the whole financials market.”
Listen at 48:54
The current market is in a manic phase of speculative excess.
“we're definitely in the the manic phase”
Listen at 51:24
The current speculative phase could continue for another one or two years.
“I think it could go on for another year, two years”
Listen at 51:54
A major AI player will eventually face financing problems.
“One of the major players will face financing problems.”
Listen at 52:47
Chinese open AI models could cause a sharp erosion in AI-industry profitability.
“Will there be competitive pressures? from, I don't know, open models produced by the Chinese that suddenly causes a mass erosion in profitability.”
Listen at 53:19
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.