The 25-year cycle: how history is pointing advisers toward small caps right now

Peter Carpi does not hedge his views. Twenty-six years at Wellington Management, one of the world’s largest investment firms with over $1.9 trillion under management, and his conviction on small caps is as direct as it gets.

Using over 100 years of public market data, Carpi maps a consistent 25-year cycle between large cap and small cap dominance, each driven by a transformative technology that concentrates returns before diffusing into the broader economy.

Every cycle follows the same pattern

History keeps telling the same story. From railways, steel and oil in the 1880s, through automobiles, radio and electric utilities in the 1920s, aerospace and consumerism in the post-war era, the nifty 50 consumer brands of the 1960s and dot-com networking in the late 1990s, every major technology cycle has produced the same dynamic: extreme concentration at the top, followed by a rotation as the technology spreads through adjacent industries and the broader economy.

Each cycle runs approximately 25 years, with large caps dominating for roughly 12 years before small caps take over for the next 12. The current large cap cycle is now 13 years old.

According to Bank of America research cited by Carpi, market concentration today has only once been more extreme in history, back in the late 1880s railway cycle, when railways, steel and oil accounted for 41 per cent of the value of public markets. Today, AI-associated names account for 40 per cent.

“We are one year past where it should have switched over,” Carpi says. The Russell microcap index has outperformed the S&P 500 in six of the last seven quarters and has outperformed by 30 per cent over the past year alone. “In my opinion, it’s already changed. We’re already in the small cap cycle.”

For advisers who felt they missed the Magnificent Seven run, Carpi’s message is unambiguous. If you are looking at investing in small caps, you are now one year into what could be an eleven-year cycle.

AI does not end at Nvidia. It diffuses through small caps

Carpi takes the cycle thesis as settled. The more pressing challenge for advisers is how to access the opportunity sensibly, and his answer lies in the companies building everything around the dominant AI names.

“Nvidia makes a chip. It goes on a board with other chips made by other people. The board’s made by other people. It’s stuck in a rack made by someone else with a bunch of power equipment next to it made by someone else, built in a data centre made by someone else, with a cooling tower made by someone else, with power behind and in front of the meter, all made by someone else.”

Every one of those businesses is a small cap opportunity. Carpi walks through several holdings to illustrate.

Solaris Energy Infrastructure, formerly an oil field services business, pivoted to providing behind-the-meter power for AI data centres on long-term contracts after Elon Musk approached them when a public utility told him grid power for his Memphis data centre would take seven years.

Centuri, an engineering construction company working for multiple utilities, is directly benefiting from the projected jump in US electricity demand growth from 2 per cent to 5 per cent annually because of data centre build-out.

NETSCOUT and A10 Networks, both unglamorous businesses by traditional measures, are becoming structurally more important as AI inference creates entirely new east-west traffic patterns inside data centres that require sophisticated network monitoring and application delivery.

“These are still just inflecting,” Carpi says. “We’ve got 11 years ahead of us.”

How global research finds what sell-side analysts miss

A case study which captures this dynamic sit with Ramaco Resources, a metallurgical coal producer in Appalachia.

Australia supplies 80 per cent of the global seaborne market for metallurgical coal. The diesel supply disruption caused by the Strait of Hormuz closure is forcing Australian mines to prioritise high-grade ore and defer preparation work, which will eventually compress production and push metallurgical coal prices higher.

Australian metallurgical coal prices have already moved from $200 to $240 a tonne in two months and are continuing to rise. Ramaco, a US producer with costs of $92 a tonne, is quietly becoming one of the most cost-advantaged producers in the seaborne market.

The punchline is that on Ramaco’s most recent earnings call, not one of the three sell-side analysts asked about Australia. They were all focused on an unrelated Wyoming rare earths project.

“This is the stuff that I love,” Carpi says. “These are examples of how we use a global research footprint to not only look at the first order of effects, but the second and third order of effects and predict what’s going to happen in advance.”

What advisers need to get right in small caps

Carpi is equally direct about where investing in small caps goes wrong. The biggest mistake is capital adequacy.

Small cap businesses, unlike large caps, cannot always access external capital during market stress or recessions. Business models that depend on continuous external funding are vulnerable precisely when conditions tighten.

Knowing the difference between a business with genuine financial durability and one chasing growth on borrowed capital is, in his view, one of the key skills that separates disciplined small cap managers from the rest.

For advisers building positions in this space, investing in small caps demands more than a strong track record. The research resources, global footprint and discipline to distinguish durable opportunities from those with a limited window are critical.

The cycle may have eleven years to run. But the returns will not be distributed evenly, and the gap between the managers who can see the second and third order effects and those who cannot is where the real performance difference will be made.