Deep Dive – June 2026

‘To infinity, and beyond!’

Reportedly a comment from the Pixar Animation Studios equivalent of Warren Buffet, Buzz Lightyear, speculating yesterday on the Nasdaq composite index!

Well, no actually, although he has been known to utter those words, I don’t think a Toy Story animated spaceman will ever be seen as offering an alternative to the Wall Street Journal. Even though the film script writers are probably just as capable as the financial journalists when it comes to guessing next week’s stockmarket valuation.

The quotation is certainly appropriate to the staggering and in my opinion completely irrational performance of the US stock markets over the last few weeks, more or less starting one month after the invasion of Iran. Readers will know that this is mainly driven by all sorts of stocks loosely connected to the booming investment in AI infrastructure. That is not the same as saying it relates to booming profits or even actual profits growth for AI companies.

I have already set out some evidence as to why the underlying fundamentals suggest that this growth in equity valuation cannot be sustained (May’s Deep Dive), and some of those facts have even got worse since I wrote about them. So why is this all happening?

The answer is, as you might suspect, something we can only speculate about until we can look back at this period as history and do a full post mortem. However, such speculation could be useful!

In this month’s article, I want to propose a comparion with the mainly UK railway companies boom (or more accurately boom years) of the mid and late 19th Century. I have previously mentioned in passing that similarities ocurred to me from my long term enthusiasm for classic English writers and now I have used the speed of AI supported research to pull together a fuller history in outline and then go on to make what I hope is an informative comparison.

What was the ‘Railway Company Stock Bubble’?

Nearly all of what follows in this specific section only is AI generated because I think for our purposes we want a lucid summary of the main events and are not overly worried about precise facts. I am therefore not concerned about over simplification or missing bits of detail in this context.

In any event, there are always multiple versions of history and many fat books analysing what supposedly happened and why. I don’t go along with the famous Henry Ford quote that ‘history is bunk’ but I do feel quite sure that one person’s version of history is quite different to another’s. Knowing now how ordinary Russian citizens of my age and their children and grandchildren think of World War 2 has been enlightening on that point!

So, here is more or less what happend and you will no doubt instantly see why on the face of it, we could be seeing history repeat itself. I will later in this article explain some differences that are not immediately obvious.

Here is the AI report (with one or two giveaway errors!) and my italics:

The British railway stock boom of the nineteenth century is usually divided into five distinct stages, culminating in the famous Railway Mania of 1844–47.

1. Pioneer Phase (1820s–early 1830s)

The success of early railways convinced investors that rail transport could be commercially transformative.

Key developments:

Opening of the pioneering (?) in 1825.
Opening of the (?) in 1830, the first major inter-city steam railway.
Early railway companies paid respectable dividends and demonstrated that passenger traffic could be highly profitable.

Investor sentiment was enthusiastic but still relatively grounded in actual operating experience.

2. The “Minor Railway Mania” (1835–1837)

The first speculative railway boom occurred during the broader economic expansion of the mid-1830s.

Characteristics:

Dozens of new railway companies were promoted.
Parliament authorised many new lines.
Railway shares rose sharply as investors anticipated nationwide network growth.

However, this boom was interrupted by financial tightening and economic weakness. Railway share prices fell heavily after 1837 and remained depressed for several years.

3. Recovery and Revaluation (1842–1844)

This was the crucial transition from ordinary optimism to full-scale speculation.

Several factors coincided:

Economic recovery after the early-1840s downturn.
Strong harvests and improving industrial output.
Falling interest rates.
Existing railways began generating better traffic and dividends than many investors had expected.

Investors increasingly concluded that railways were the infrastructure of the future.

Rather than being viewed as risky experiments, railway shares came to be seen as a near-certain route to wealth.

4. The Great Railway Mania (1844–1846)

This was the classic speculative bubble.

Features included:

Railway share prices nearly doubled between 1843 and 1845.
Hundreds of new railway schemes were promoted.
Investors subscribed to companies that often had no track, no land and sometimes only sketchy plans.
Shares could often be purchased through instalments, increasing leverage and speculation.
Parliament was overwhelmed with railway bills; more than 700 were considered in 1846.

The peak came in 1845–46:

263 railway Acts were passed.
Roughly 9,500 miles of proposed routes were authorised.

At this stage, investors were no longer mainly valuing existing profits. They were pricing in enormous future growth and assuming that virtually every proposed line would succeed.*

*NOTE THAT LAST AI GENERATED LINE – it sounds familiar does it not?

5. Collapse, Consolidation and Aftermath (1846–1850s)

The boom ended when reality caught up with expectations.

Several pressures emerged:

Interest rates rose.
Capital became scarcer.
Construction costs proved enormous.
Many authorised lines were commercially unviable.
Investors struggled to meet instalment payments on their shares.

Share prices first stagnated and then entered a prolonged decline. By the late 1840s many railway companies had failed or were absorbed by stronger rivals.

Yet unlike many speculative bubbles, the physical infrastructure largely remained:

Thousands of miles of railway were actually built.
Britain acquired the foundations of its national railway network.
Large systems eventually emerged through mergers and consolidation.

One-line timeline

Stage	Dates	Character
Pioneer railways	1825–1834	Proof that railways work
Minor Railway Mania	1835–1837	First speculative boom
Recovery & revaluation	1842–1844	Rising profits and optimism
Great Railway Mania	1844–1846	Full speculative bubble
Crash & consolidation	1846–1850s	Share-price collapse, network survives

A striking feature of the railway boom is that it was simultaneously a financial bubble and a genuine technological revolution. Investors lost large sums, but the capital raised helped create much of the railway system that powered Victorian Britain’s industrial economy for the rest of the century.

That’s the end of the AI content.

Elements that are similar to the AI boom

We are talking about genuine technological developments that started slowly and then appeared likely to be widely adopted.
There was a requirement to build out a vast amount of expensive infrastructure, soaking up capital, well before the real income flows, profits and dividends came on stream.
Any number of businesses were formed and floated at speculative valuations with prospectuses announcing projects that not many investors wanted to, or were even capable of understanding.
Investors entered the market who otherwise would not have participated, notably those not generally ranked as wealthy and a number would have borrowed money to participate.
It was not a mistake to believe that the world would be different after the new technology was fully mature and in everyday use. Anyone who loves Sherlock Holmes stories will know how a comprehensive and carefully scheduled railway service across England allowed Holmes and Watson to flit between crime scemes and Baker Street, probably faster than we could today!
The railway expansion soaked up natural resources like land and coal, just as the AI boom requires vast amounts of water and electricity.
In time competition between the main initial market entrants sorted out the winners and losers with a few large companies carving up the market, mainly by geography with the railways, more likely by consumer type with the AI services.
Politicians actively participated, promoting companies and lining their own pockets.
Finally, I would venture that human nature has not changed, even if we now have more labels from psycholgists to help explain it.

Readers may be able to think of other parallels.

What is different?

The structure of companies was in the early part of the 18th century such that shareholders had unlimited liability – they actually were the company, so they took not only dividends (if any) as a share of profits, but also carried personal liability for the company’s debts.
Shares were more commonly issued part paid (still possible now) with instalments due that not every would be millionaire could actually come up with when asked.
Today, there are many ways of betting on shares that do not involve direct ownerhip, like derivatives contracts and the prediction markets.
There are millions of investors now who will buy shares without any personal decision, and in increasing quantities as companies swell in terms of market capitalisation, because of the huge flows of money into passive index funds.
Regulation 150 years ago was of course much lighter – when we look back in time regulation was always lighter in the past! It is usually empty stables that get their doors bolted when it comes to financial regulation.
The underlying economy was very different in the 19th century. In essence, the class system was still largely intact and a wealthy business class was newly emerging. Large amounts of long term capital were tied up in landed estates and Gilts. Later in the century capital was widely deployed overseas as the concept of ’empire’ developed.
Unitised or fund type investments were not available. Interestingly, from 1868, there was available the Foreign and Colonial Investment Trust, the grandaddy of the collective investment world.
Dealing in shares involved multiple specialist professionals, all taking a cut, and settlements took weeks as opposed to seconds.
Most importantly, in the 19th century information was limited to printed material and it was likely difficult to assess its reliabilty, although literature from the period suggests many readers of newspapers took them to be ‘gospel’. Checking on investments overseas generally required people actually undertaking often hazardous sea passages and risking disease in unpleasant climates. I suppose one might suggest that visiting the USA is hazaradous in some ways now and being ill there would be very expensive, but as a rule such travel is no longer essential!

What lessons can we take from the past?

To quote another old chestnut, allegedly from Mark Twain: ‘the past does not repeat itself but it merely rhymes’, or similar words. The essence of this observation is that human behaviour is to a degree pre-programmed, so given a similar environment, threats or opportunities and comparable resources, people tend to follow a pattern of activity.

It is difficult to claim that humanity typically learns from its mistakes and does not therefore repeat them when you think about the USA entering 8-10 wars (depending on definition) in the Middle East over the last few decades, or the number of times that people have died after governments relaxed safety standards, or how often huge amounts of taxpayers money are put in the hands of bloated self-preserving bureaucracies. And so on.

So what might we expect to see happen as the AI boom progresses, if we take ‘railway mania’ as the model?

Firstly, that any sharp upturn in share prices might be sustained for several years, however unjustified that is by normal valuation metrics.

Secondly, that businesses with no real advantages, not even a real product, will get valued at incedible prices just because they can in some vague way be associated with AI developments.

Thirdly, that large amounts of money will be borrowed to get in on the action, introducing massive gearing rsik.

Fourthly, that goverments will want to be seen to be keeping up to speed, even if they actually are not, and will invest in projects, even creating their own and in most cases they will do so unsuccesfully.

Fifthly, that a small number of companies will make very large profits, with it being different companies at different stages, and an even smaller number will earn continuous solid profits and pay growing dividends for a period, but even for them, that period may not be very long.

Finally, that there will be a collapse, that it will be dramatic and will wipe out quite a number of people and businesses financially, but that a part, possibly the greater part, of the new infrastructure will be useful and quite likely utilised by a different set of businesses than those that financed it.

One might also speculate that there could be ‘echos’ of the cycle down the road as the technology evolves and its application becomes clearer. Possibly, there may even be a contraction and repurposing of data centres. I would not be surprised if in less than 10 years time people ran their AI models on their smartphone and hardly ever accessed the internet for basic AI daily use.

What actions might make sense?

I would suggest a retail investor, not wanting to be the candidate for a modern Dickensian tale of financial woe, has 3 choices:

Avoid AI boom stocks as far as possible. Actively exclude the risks of high volatility. Trying to time the market to pick up the momentum and exiting before the crash is only for the very skilled, very brave or very foolish.
Invest in one or two funds that are narrowly focused on properly underpinned growth companies, a few of which may be beneficairies of AI infrastructure spending (like ASML, TSMC, Intel and so on). This option focuses on trusting the fund manager to follow the momentum, but with care.
Ignore the AI boom completely as a factor in asset allocation and fund selection decisions, but assume that the fund managers chosen will make the right decisions for you and that may or may not include joining the speculation. This is the ‘agnostic’ or ‘hands-off’ course.

Note that if you like passive index tracker funds from Vanguard and the like, or ETFs built on specific indices, you won’t have any choice but to own the boom stocks at the heavy market capitalisation weightings they will inevitably enjoy. This is an increased risk in my opinion.

As ever, I welcome questions and comments on all of the above.