"*" indicates required fields
or call +44 (0)1603 672777
Market insights
Director
Richard’s reflective investment commentary links a personal memory of craftsmanship to today’s AI-driven market boom. This article compares the current surge in AI investment with the railway boom and Panic of 1873, exploring both the similarities- transformative technology, heavy capital spending and speculative enthusiasm- and the key differences, including stronger corporate balance sheets, faster adoption and greater financial-system resilience.
Laurence Bernard Bede Brooks was born at the tail end of the 19th century and died just sixty-six years later, when the heart that had been irreparably broken two years earlier by the loss of his only daughter gave up. He was a gifted painter and a skilled cabinetmaker. His council house was one of the Homes for Heroes he’d helped to build after returning from the horrors of the First World War. He hand-crafted the furniture that filled the home he lived in for the rest of his life. He was my grandfather, and I still have some of his pieces.
His shed was where my most treasured memories were forged. I was just eight when he died, but his lessons have stayed with me: studying the grain to select the best piece of oak, making the saw do the work, sharpening his carving chisels before returning them to the oilcloth wrap. And the smell, most of all I remember the smell of wood shavings and turpentine, linseed and oil paint.
Not long ago, in a back alley in Dubrovnik’s Old Town, we came across a small workshop where an artisan churned out tiny wooden boats for tourists searching for authenticity. The workshop was unattended, but the door was open, so we wandered in. The smell was that same mix of wood shavings and turpentine; closing my eyes, I was back in the shed. I could almost hear my grandfather.
Of course, allowing our minds to be transported back isn’t always evocative of better times. For the last couple of years we’ve seen strong market returns; close your eyes, and you can be quickly carried back to any of a multitude of financial booms and busts. Markets are like children: when they get over-excited it’s only a matter of time before their exuberance melts into tantrums and tears. And there are growing concerns that the AI boom fuelling much of the present market recovery is edging into sugar‑rush territory.
I’ve been reading about the Panic of 1873, which triggered a brutal, multi‑year economic contraction known as the “Long Depression” (which originally held the title of the “Great Depression” until the 1930s). It transformed the global economy, shifting the balance of industrial power and reshaping financial systems across the US and Europe. The parallels and contrasts with today’s markets are striking.
1. A revolutionary technology at the centre
The early 1870s were the height of the railways boom, the AI of their age. Railways fundamentally changed how goods, people and information moved, just as AI is changing how knowledge work is performed. Investors believed both technologies would reshape the economy, and they were right.
2. Massive capital investment
Railway construction required vast amounts of capital for tracks, land, bridges and rolling stock. Similarly, AI requires enormous spending on chips, data centres, energy infrastructure and software development.
3. Speculation outrunning reality
In the early 1870s, investors financed more railway capacity than the economy could immediately use. Many lines were built before demand fully materialised. Likewise, today there is debate over whether AI infrastructure spending is running ahead of actual commercial returns.
4. Concentration of investor enthusiasm
During the railway era, capital became concentrated in a relatively small number of high-profile railway projects and financiers. Today, market performance is heavily influenced by a small group of AI-related firms, especially Nvidia, Microsoft, Alphabet, Amazon and others.
5. New investors entering the market
Both booms attracted investors who were investing in the theme as much as the underlying economics. Railway securities were sold broadly to retail investors; AI-related investments have attracted retail investors, venture capital and institutional money eager not to miss the next big thing.
6. “General Purpose Technology” characteristics
Railways reshaped many industries, from agriculture to manufacturing. AI appears to similarly transformative across professional services, healthcare, finance, manufacturing and education.
1. Today’s leaders are already profitable
One of the biggest differences is that many leading AI companies generate enormous profits and cash flow today. By contrast, many railway companies relied heavily on external financing and carried substantial debt long before their economics were proven.
2. Infrastructure versus applications
Railways required physical construction before benefits emerged. AI software can be deployed globally almost instantly. This is likely to shorten the lag between investment and productivity gains.
3. Financial system resilience
The 1873 system lacked a central bank capable of acting as lender of last resort. The collapse of Jay Cooke & Co. helped trigger widespread banking panic. Modern central banks, deposit insurance and regulatory systems provide far greater shock absorption.
4. The technology already works
By 1873, railways were being built ahead of demand, with doubts around many routes ever becoming profitable. In AI’s case, the technology already demonstrates real utility in coding, research, customer service, content creation and knowledge work. The debate is less about whether AI works and more about how much value will ultimately accrue to investors.
5. Speed of adoption
Railway network effects took decades. AI adoption has been extraordinarily rapid – generative AI has reached adoption milestones faster than either the PC or the internet.
6. Global information flow
Investors in 1873 operated with limited information and slow communications. Today’s markets receive real-time information, corporate disclosures and performance data, making capital allocation more efficient, though not necessarily immune to bubbles. The collapse of the 1870s was exacerbated by widespread fraud; today’s improved governance, transparency and speed of communication make this less likely
Today’s market faces an echo of 1873’s core vulnerability: the risk of an over‑leveraged bet on future productivity. If infrastructure spending overshoots reality, or if geopolitical energy shocks severely compress consumer purchasing power, certain overextended sectors will face a harsh reckoning. None of those risks feels far‑fetched.
Yet, because today’s corporate balance sheets, particularly in tech, are anchored by immense liquidity rather than fragile bank debt, the threat is more likely to manifest as a painful valuation adjustment and slower growth than a systemic shutdown.
While a correction should always be seen as an inevitability rather than a possibility, when it comes, it is unlikely to be overly damaging and likely to be confined to a small number of sectors. Indeed, as Katie Martin observed in a recent Financial Times article, we are already seeing this playing out with attention moving from the Magnificent 7 big AI spenders to the chip manufacturers, on average the Mag 7 stocks are flat this year while the Philadelphia Semiconductor Index was up 80% over the first half, vindication, for now at least, of our policy to capture total market returns rather than trying to pick the winners.
The boom in AI expenditure is effectively a private‑sector fiscal stimulus. It goes some way to explaining the apparent resilience of markets and economies throughout a series of geopolitical and policy‑driven shocks.
And what if AI ultimately disappoints? I’m reminded of Keynes’ thought experiment where he showed that paying one group of unemployed men (it was before women were expected to wield a shovel) to dig a hole and another to fill it in again resulted in a net economic gain (through things like the multiplier effect), even though the activity itself was pointless. On that basis, in the short-term all the activity involved in building data-centres and chip fabs will boost the global economy even if in the long-term it all ends in disappointment.
There was one lesson that sadly failed to stick. At the end of every day in the shed, the last fifteen minutes were spent tidying and sweeping up. As anyone who has seen my desk will attest, that’s not a habit I developed.
Richard Ross July 2026
This blog post is intended for information only and does not represent personal financial advice. If you would like to speak to one of our Chartered Financial Advisers, please get in touch. Past performance is not a guide to future performance. The value of an investment can fall as well as rise and is not guaranteed; you may get back less than you paid in.