Are AI stocks the new railroad bonds?

Summary

Podcast Summary: Unhedged – "Are AI stocks the new railroad bonds?"

Date: May 5, 2026
Hosts: Katie Martin (KM), Robert Armstrong (RA), Robin Wigglesworth (RW)
Produced by: Financial Times & Pushkin Industries

---

Overview

In this episode, the Unhedged team explores a timely and thought-provoking question: Are today’s AI stocks the modern equivalent of 19th-century railroad bonds? The hosts use their deep knowledge of financial history to dissect current market dynamics, drawing direct parallels to past technological booms and busts—most notably, the US railroad expansion and the dot-com era. Along the way, they debate whether we can (or should) learn from history, the lasting impact of speculative bubbles, and how today’s geopolitical uncertainty and tech hype might shape the legacy of our financial age.

---

Key Discussion Points & Insights

The Value of Financial History

[00:44-02:51]

• Why study financial history in the first place?
  • RW admits his personal fascination: "I do it because I love it. I just think it's fascinating writing." (RW, 02:11)
  • Learning from the past is critical, though not always straightforward: “Those that learn from the past usually kind of repeat some of the screw ups in the future.” (RW, 02:24)
  • “Those who do know the past are also doomed to repeat it.” (RA, 02:38) — highlighting the limits of even well-meaning historical parallels.

The Railroad Boom: Scope, Bust, and Lessons

[03:51-07:58]

• How big was the railroad boom?
  • RW contextualizes the scale: “It was a gargantuan buildup… even the AI boom today looks like a tiny little gnat on the ass of an elephant compared to the railway boom.” (RW, 03:56)
  • By GDP-adjusted terms, the US railroad bond boom equals about "$10 trillion today." (RW, 04:30)
• Railroads as miraculous technology:
  • Public perception was awe and fear (“People were worried that people would essentially disintegrate at speeds over 30 miles an hour.” RW, 05:14)
• Aftermath: The Long Depression
  • Multiple global busts; the 1873 crisis triggered by European liquidations: “There was an epic one… in 1873. It was probably one of the first big global financial crises.” (RW, 06:53)
  • Terminology shift: “It was called the Great Depression until the actual Great Depression made this just known as the Long Depression.” (RW, 07:45)

Parallels to Today’s AI Boom

[08:00–13:14]

• Technological utility vs. investor pain:
  • “The technology itself can be super, super, absolutely indisputably useful, but you can still end up losing a ton of money on it.” (KM, 08:54)
• Bubbles aren’t all bad:
  • “The optimal level of bubbles and busts is not zero... That’s sort of part and parcel how we move forward as well.” (RW, 09:37)
• Telecom bubble as a more recent example:
  • “Most people involved got crushed… but the result was we got this wonderful productive asset which we're still using today…” (RA, 09:51)
  • The open question: “Are these AI data centers long-lived productive assets?” (RA, 10:40)
• Hopeful outcomes:
  • Even if AI doesn't pan out, built energy infrastructure may still prove useful. “Hopefully that may be… useful. So I’m optimistic, though I suspect… people will lose an ungodly amount of money.” (RW, 10:56)
• The Apple contrarian strategy:
  • “Apple is saying we’d rather rent than own and we’re going to see what happens here.” (RA, 11:35)
  • “There is somebody out there… looking at the example of the railroads… and saying we’re going to take a little bit of a wait and see here.” (RA, 11:43)
• Major difference from the past:
  • The AI boom is led by “vastly profitable companies… even if… they lose every single dollar now putting into AI, these companies… aren’t for the most part going to go bust.” (RW, 12:37)

The Uncertainty of Historical Parallels and Defining the Era

[13:14–16:26]

• Reflecting on the present moment:
  • “What are they going to call this period… in 2050? …There is this AI boom. Whether it busts or not, none of us know. And we should be honest about that.” (KM, 13:57)
  • Geopolitical uncertainty: “There is what clearly seems to me to be like an unraveling of the global geopolitical order… things like NATO and the UN are under huge amounts of pressure.” (KM, 14:26)
  • Shifts in attitudes toward the dollar, US bonds, and markets.
• What will define this financial era?
  • RA: “If… that [post–Covid inflation] was just the first tremor in what turns into a higher inflation, higher interest rate era, that is what we will be remembered for.” (RA, 14:54)
  • RW: The Iran war could be defining, especially if oil crises erupt. “If we do… large parts of the world run out of oil, then suddenly… this might define a longer time period.” (RW, 15:47)
  • Both agree: The AI boom's true impact will only be visible in hindsight—"whether it's a disaster or a huge benefit… we're at the foothills of it right now" (KM, 16:26)

---

Notable Quotes & Memorable Moments

• “Even the AI boom today looks like a tiny little gnat on the ass of an elephant compared to the railway boom.” (RW, 03:56)
• “People were worried that people would essentially disintegrate at speeds over 30 miles an hour.” (RW, 05:14)
• “The optimal level of bubbles and busts is not zero… we almost need these [so] we need people to lose the senses to do these big build outs.” (RW, 09:37)
• “If the money that goes bankrupt is spent on a long-lasting productive asset, it can be a good thing in the long run for an economy.” (RA, 10:03)
• “What are they going to call this period… in 2050? …There is this AI boom. Whether it busts or not, none of us know.” (KM, 13:57)
• “The moral of the story is things that seem to be forces of nature and make perfect sense today, might not make perfect sense tomorrow…” (KM, 13:14)
• “Even if… they lose every single dollar they are now putting into AI, these companies… aren’t for the most part going to go bust.” (RW, 12:37)

---

Important Segment Timestamps

• [02:11] – Why financial history matters (RW)
• [03:56] – Size and scope of the US railroad boom (RW)
• [06:53] – The 1873 crash and its global impact (RW)
• [09:37] – Productive bubbles and the telecom example (RW/RA)
• [10:40] – Lifespan & impact of AI data centers (RA)
• [11:35] – Apple’s counter–AI build-out approach (RA)
• [12:37] – Key difference: 21st-century tech's profitability (RW)
• [13:57] – Defining the current era: Inflation, geopolitics, AI (KM/RA/RW)

---

Long/Short Segment

[17:33–18:15]

• Robin (RW): Short Jamie’s deal to try to take over eBay
  • “It is a mouse pumped full of medical cocaine trying to eat a massive elephant with a highly confident letter from a Canadian investment bank. This deal is not going to happen, but it's going to be highly entertaining for us to follow it.” (RW, 17:57)
• Katie (KM): Long Tiny Tiny Ice Creams (Magnum’s move to small, portion-controlled ice creams aligned with weight loss trends)

---

Key Takeaways

• Financial history provides context for today’s tech/AI boom, but doesn’t offer a crystal ball.
• Even world-changing technology (like railroads) can trigger destructive busts, though productive assets may endure.
• There’s value (and risk) in speculative bubbles—sometimes bankruptcies lay groundwork for future growth.
• Today’s AI build-out differs thanks to the cash flow strength of leading companies.
• Whether the “AI era” will define a generation is only answerable in hindsight; inflation and geopolitics are major wildcards.

---

This summary provides a structured, detailed recap of the discussion, capturing the lively and incisive tone of the Unhedged podcast. For readers looking for both investment insight and perspective, the parallels drawn between past and present are especially valuable—and delivered in classic FT fashion, laced with both seriousness and dry wit.

Transcript

[A]

Bull or bear, Trade or tariff, Future or fad? There's more than one side to every story. With the Flipside podcast from Barclays Investment bank, you'll hear two research analysts having a provocative debate on hot topics in business and markets. Listen to the flip side on your favorite platform,

[B]

Pushkin. You want to figure out what financial markets are up to, what they're trying to tell us. The go to way that professionals try and do that is by looking not at the future, but at the past. So if you look at the AI boom in markets today, the big question is whether this is like the dot com boom and bust of 25 years ago. This game is less about gazing into a crystal ball and more about being low key. A bit of a history nerd. So today on the show, why it pays to know your financial history. This is Unhedged, the Markets and finance podcast from the Financial Times and Pushkin. I'm Katie Martin, a markets columnist at the FT in London where we have a wobbly bond market and a local election coming up. What can possibly go wrong? Joining me today, I have a twofer, ladies and gents. Yes, I have that big Rob Armstrong over there in New York. Rob, say hi.

[C]

Hello.

[B]

Don't put on a funny English accent, Rob, because we also have an actual funny English accent. An actual funny English accent from Robin Wigglesworth who's the editor of FT Alphaville and also a card carrying financial history nerd over there in sunny Oslo. Robin, hello.

[A]

Hello. Should I do a Norwegian accent today then?

[B]

No need for that. So, Robin, I know you are loathe to indulge in shameless self promotion, but you've been deep in the weeds on financial history lately. The thing you're doing at the moment is this podcast also in the FTSE stable with Gillian Tet, about the story of money, which is about financial history. Why would you do such a thing? Who cares about the deep history of finance? Why is this important?

[A]

Well, I do it because I love it. I mean, I'm, as you repeatedly point out, a massive nerd about these things. So I just love it for its own sake. I just think it's fascinating writing. But like you say, we do learn a lot from the past. I, you can choose the George Santayana quote or the, the Winston Churchill quote or even go back to Edmund Burke. But you know, those that, that learn from the past usually kind of repeat some of the screw ups in the future.

[C]

I've always thought that this quote was absolutely true. But those who do know the past are also doomed to repeat it so

[B]

well, that was like.

[C]

So that's a little bit of a problem.

[B]

But bringing this somewhat more up to date, like the thing that is really driving stock markets at the moment. Stock markets have decided that the Iran war is like old news and they are just like plowing on regardless. And one of the reasons they're plowing on regardless is that they are really tightly glued to corporate earnings. And really big companies, particularly kind of AI flavored tech flavored companies, are just making money hand over fist. But that sort of links back to this sort of gnawing thing that people have in the back of their head, which is, is this a little bit like the dot com boom and bust of sort of 1999, 2000 or worse? Is this like the US railroad boom and bust which happened sort of from the, the late end of the 19th century? And I guess you can draw a thread from there through to the proper Wall street crash of the 30s. You've been talking about this on, on the POD lately. What like, do you think there are. Well, first of all, for people who don't know, what was the US railroad boom and bust?

[A]

Well, essentially in the 19th century, there was this gargantuan, and I mean it, you know, that's, I'm not exaggerating. It was a gargantuan buildup of railways around the world. It was, you know, even the AI boom today, it looks like a tiny little gnat on the ass of an elephant compared to the railway boom. So, you know, I saw some numbers from Morgan Stanley. They think that the hyperscalers are going to spend maybe a trillion dollars on data centers by 2027. Now, but the railway boom just in the US just on the bond issuance, there was around 5, 6 billion dollars worth of bonds issued. And that doesn't sound like much, but if you scale it relative to the size of GDP at the time, because the US was a small economy, that's the equivalent of $10 trillion today.

[B]

Wow.

[A]

So the railway boom, the bond issuance, and like some of this was financed by stocks and government grants and so on, but largely by bonds, was around 10 times the size of today's AI

[C]

give us time, Robyn.

[A]

No, no, exactly. Fingers crossed. We get there.

[B]

The week is young. But like, at the time, was there much evidence of like, you know, the, the olden times equivalent of miserable people like me saying this all seems a bit rid.

[A]

Well, you have to remember that the railway was considered a miraculous technology when it first arrived in the uk People were worried that people would essentially disintegrate at speeds over 30 miles an hour. People generally thought that, like humans weren't built for those kind of unimaginable speeds. We just turned to dust or at the very least go mad at around 20, 25 miles per hour.

[C]

And so it wasn't only considered miraculous technology, it effectively was. I mean, this was. It was. Yeah, yeah, absolutely amazing. The things had changed, right?

[A]

I mean, the US was the epicenter of it. I mean there was. People went railway gaga everywhere in the world. And there were railway booms and busts multiple over the the 19th century. But the US discovered because of its scale, right, it was the biggest. It was, you know, where they went the most crazy for it. And it my. I'd argue they literally made the United States. The United States were not united until the railways came. And this was a country that basically had a bit of like west coast stuff, a big chunk of the middle that was just frontier country and then. And it ran along a north to south axis. But after the railways, you could suddenly buy like Maine lobster in California or Californian apples in New York. You could literally travel across the country. It's something that was unimaginable. It was like Apollo. Well, it's really like 20 or 30 Apollo programs, essentially.

[C]

So essentially people issued a third of GDP in debt, the equivalent of that. How did that all turn out?

[A]

Well, it didn't turn out great for all the bondholders. You know, spoiler alert though you probably see it coming. There were multiple busts along the way. There was an epic one. I just like a humdinger in 1873. It was probably one of the first big global financial crises because it arguably was triggered by a crash in Austria that ricocheted over to the US or basically meant that lots of European investors lost a ton of money in Austria and they had to liquidate their bonds. There were massive investors in American railway bonds. So these were some Dutch burgers, German nobles, British merchants, and they started liquidating bonds in the U.S. and then just everything just collapsed at the U.S. so this was. It was called the Great Depression until the actual Great Depression made this just known as the Long Depression.

[C]

Yeah, it's like how the, the Great War. The Great War became World War I after World War II came along.

[A]

Yeah, exactly.

[C]

Yeah, yeah, exactly.

[B]

So, like, it strikes me there's two like, takeaways from there. One is that everything is connected. And so, you know, a sell off in one particular asset in one particular country may not have any logical link to another asset class in another country. But Once one thing starts selling off and people start taking losses, they end up having to sell stuff elsewhere to stay current on their bills. And that is still one of the defining features of global markets today. The other is railroads are quite useful. They help you get your apples and your lobsters and all that sort of thing. The technology itself can be super, super, absolutely indisputably useful, but you can still end up losing a ton of money on it. And that I guess is the debate that's going on in AI. Now did that feel a little bit, kind of almost like sort of prescient and spooky to you while you were reading up on all this?

[A]

It did, because it's, you know, sometimes you can go a little bit crazy seeing parallels in the past, but with AI, they're slapping you in the face. It's, it's very obvious, it's quite similar. Even though the technologies are different, you know how people think they can be transformative. The physical capex buildup, because some financial bubbles are almost purely financial, this was a real one. I do think that, you know, bubbles and bursts get a bad rap that the reality is that the optimal level of bubbles and bus is not zero, that we almost need these, we need people to lose the senses to do these big build outs. And yeah, people do lose money and unfortunately sometimes there are individual tragedies or even societal tragedies among it, but that's sort of part and parcel how we move forward as well.

[C]

A terrific example of this Robin, is the lesser known bubble of the early 2000s, which was the telecom bubble of the early 2000s. And people spent an epic amount of money putting fiber optic cable in the ground. A lot of companies went bust every, you know, most people involved got crushed one way or the other. But the result was we got this wonderful productive asset which we're still using today, which is tons of fiber connections, you know, and tons of telecom infrastructure. And there's, there's loads of real estate stories like this, there's the railroads, etc. Etc. That these booms, if, if the money that goes bankrupt is spent on a long lasting productive asset, it can be a good thing in the long run for an economy. The question is, are these AI data centers long lived productive assets? They're certainly going to be productive. How fast they depreciate all that stuff is, is another question altogether.

[A]

Yeah, I'm hopeful that the, the energy related infrastructure buildup is going to be helpful because obviously you can't just build a data center, you need to power it as well, and hopefully that may be, you know, if, if AI doesn't quite work out that it, you know, doesn't quite revolutionize the world. We don't need all these data centers. Well, we've built a lot of energy infrastructure and renewable energy and new gas turbines and things like that that you know, will be useful. So I'm optimistic though, I suspect, you know, people will lose an ungodly amount of money at some point. I hope still 20 years on will maybe say that maybe it was overall worth it for the world.

[C]

Well, it's interesting Rob, and I'm writing right now a piece about Apple which is kind of the company that is taking the other side of the AI bet Famously all the other huge tech companies, all the other magna, most of the other magnificent seven tech companies are loading money into the AI data center build out and Apple is saying we'd rather rent than own and we're going to see what happens here. So there is somebody out there, I think, looking at the example, a big money player looking at the example of the railroads or similar examples and saying we're going to take a little bit of a wait and see here and it'll be interesting to see whether that's a good bet or not. But I think it's a big bet.

[A]

I mean it could be a profitable one. Just look at how Apple works with, with Google, right?

[C]

Yes.

[A]

I mean this deal came under regulatory pressure but you know, Apple didn't build its own search engine. It basically kind of got money from Google to make Google search the default. Maybe it'll do the same with Anthropic or OpenAI. I think there are many ways people can make money out of this. Yeah. And I have to make the big difference, of course even with the railway boom of the, of the 19th century is that, you know, there is an awful lot of debt issuance happening, but by vastly profitable companies. And these tech companies still print money. So even if like they lose every single dollar they are now putting into AI, these companies aren't for the most part going to go bust. Which makes me slightly less worried than maybe I would have been if I'd been a financial journalist in 1872. Looking at the railway boom.

[B]

The moral of the story is things that seem to be forces of nature and make perfect sense today, might not make perfect sense tomorrow, and ways that look like, you know, you're shooting fish in a barrel, this is a dead easy way to make money. You look back on it in a few years time and you think what the hell was everybody thinking? So it's thinking of, you know, unhedged podcast in the year 2050. Right. It's all done by, you know, people who haven't been born yet or AI or some other thing. And they're looking back at this moment in time that we're living in right now in 2026. And what are they going to call this, this period that we're going through now? Because there's a lot of stuff going on. There is this AI boom. Whether it busts or not, none of us know. And we should be. We should be honest about that. But there's the situation in Iran. There is what clearly seems to me to be like an unraveling of the global geopolitical order. Multilateral organizations are falling apart. A post war order is falling apart. Things like NATO and the UN are under huge amounts of pressure. Things that we've all grown up with that we think are just sort of natural phenomena that come out of the ground. They are all actually quite complex organizations that are under strain. And I know that Rob will disagree on this, but there is this push among global investors to think differently about the dollar, differently about the rate profile of US Government bonds, differently about the US Stock market. What do we think is going to be the thing in the unhedged podcast of 2050 that actually sticks and turns out to be a meaningful moment? Or is it none of them? I don't know. Rob, you can go first.

[C]

I think much depends on whether we have another serious inflationary incident. So we had this big burst right after Covid of inflation, and inflation is like, you know, Robin might disagree, but I think inflation is so often the defining factor of a financial era. And if, if that was just the first tremor in what turns into a higher inflation, higher interest rate era, that is what we will be remembered for, is when inflation returned after 50 years in the wake of the crisis. What do you say, Robin?

[A]

I'm tempted to ask, like Ryan Cohen did GameStop CEO yesterday in the interview where he said that we will see was his answer to other things.

[C]

This is not allowed on the Unhedged podcast.

[A]

No, unfortunately, we are unhedged here. Yeah. Yeah. The Iran war feels like it's now on the cusp to be sort of an apocal thing. If we do basically large parts of the world run out of oil, then suddenly, I'm willing to say this will go from something that was a defining thing for 2026 as being something that might define a longer time period. Yeah, I think, definitely.

[C]

I mean, that's that's if AI turns out to be what some people say it's going to be, this year will be remembered as the foothills of AI before AI changed the economy completely.

[B]

Yeah, whether whether AI is is a disaster or whether it's a huge benefit for the world, either way, we're at the foothills of it right now. What I'm hearing here is that we all need to understand history, but that doesn't necessarily help help us to figure out what's going to happen next. Nonetheless, an interesting intellectual exercise. Listeners, we're going to be back in just one sec with Longshaw.

[A]

Today's markets move fast. Get the insights you need in 10 minutes with the Barclays Brief, a new podcast from Barclays Investment Bank. Through sharp dialogue and scenario based analysis, our leading experts analyze key market themes each week. So whether you're managing a portfolio or leading a business, the Barclays Brief podcast can help you make smarter decisions today. Stay sharp, stay briefed. Find Barclays Brief wherever you get your podcasts.

[B]

Okey doke. It is time for Long Short. That part of the show where we go long a thing we love or short a thing we hate. Robin, since you came along today to help us out, what you saying?

[A]

I'm going to be short Jamie's deal to try and take over ebay.

[B]

Oh that. Oh that is not gonna work.

[C]

That is the mouse eating the lion and it is not gonna work.

[B]

This is the Gamestop Yahoo stock.

[A]

It is a mouse pumped full of medical cocaine trying to eat a massive elephant with a highly confident letter from a Canadian investment bank. This deal is not going to happen, but it's going to be highly entertaining for us to follow it.

[B]

Yeah, it is going to be entertaining. So yeah, that is definitely in its favor. Well, I am Long Tiny Tiny ice creams so shares in Magnum Ice Cream rose quite a lot last week after it said it was shifting to bite sized portion controlled ice creams. For all those people on on the weight loss drugs who don't want large ice creams. Now we can all enjoy tiny ice creams instead. I kind of like that. Little tiny ice creams to keep us happy through the course of the day. Listeners, we're going to be back in your ears on Thursday, so listen up then. Unhedged is produced by Jake Harper and edited by Bryant Urstadt. Our executive producer is Jacob Goldstein. Cheryl Brumley is the FT's global head of audio. Special thanks to Laura Clark, Greta Cohn and Natalie Sadler. FT Premium subscribers can get the unhedged newsletter for free and a 30 day free trial is available to everyone else. Just go to FT.com unhedged offer I'm Katie Martin. Thanks for listening.