Megatrends for Equity Returns | Pasi Havia | Neuvottelija 416

Summary

Who he is. Pasi Havia manages United Bankers’ UB Megatrendit equity fund, which started on 6 March 2026; he joined United Bankers in January 2026 after more than a decade at Helsinki Capital Partners. He has managed funds professionally for about twelve years and already used megatrends in his previous fund.

Four megatrends. More than 60 per cent of the fund sits in AI and digitalisation, deliberately overweighted because that is where Havia sees the biggest return potential right now. Electrification and the energy transition is the second-largest theme, and population ageing and European defence take about 10 per cent each. The weights are a view, not a fixed allocation: Havia can shrink AI and grow less correlated themes if the cycle turns. In its first seven months the fund returned more than 26 per cent after fees.

Machines as the customer. Havia’s central idea is that demand for compute and memory no longer comes from people but from machines, “and they are insatiable”. Tokens were first consumed by people, now mostly by AI agents, and next by physical AI and edge devices that must process locally because the latency to a data centre is too long. He calls this Sami’s “nasty megatrend number one”: if machines replace human labour, they also replace human consumption.

The Micron paradox. Micron, Havia’s favourite holding, has risen about 500 per cent in a year while its multiples have stayed where they were, because the market keeps pricing it at the peak of the cycle. Havia points out that the company itself expects memory to be tighter in 2027 than in 2026, has no visibility on when supply and demand balance, and is signing long agreements with customers out to 2031. He chose an American memory maker over Samsung and SK Hynix to limit political risk.

Laser-precise exposure. Within AI he buys the narrowest points of the value chain: Broadcom as the designer of Google’s and Amazon’s custom chips, ASML and BE Semiconductor from Europe, and Lumentum rather than Nokia or Coherent for data-centre optics. Only profitable companies qualify, so SpaceX and Anthropic stay out. Multiples, he argues, break down under exponential growth: Nvidia’s P/E fell from about 100 to about 30 and Palantir’s from 600–700 to about 150 as earnings grew. Sami counters that a 5 per cent earnings yield now competes with a returned risk-free rate, and that governments compete with trillion-dollar companies for capital. Both accept that circular financing in the AI chain cannot be avoided, and Havia expects value investing, after almost 20 years of underperformance, to come back brutally one day.

Defence, energy and ageing. In defence Havia avoids state-owned companies and holds TKMS and Theon International, citing NATO’s 5 per cent target, Europe’s underinvestment and a possible Taiwan crisis around 2027–2028. In energy he covers America’s ageing grid layer by layer with Constellation Energy, Quanta Services, GE Vernova and Vertiv; Sami’s AI agents had picked Nexans instead. In ageing he prefers Eli Lilly to Novo Nordisk, partly for political risk. He does not invest in China, because Western owners do not get to benefit from the returns, as his earlier holding Huya showed overnight. The episode closes with Europe’s regulatory handicap, Finland’s case as a data-centre country and Sami’s plan to run local AI models for next year’s challenge. Nothing in the episode is investment advice.

Watch

The recording lives on the Neuvottelija channel: Megatrendeillä osaketuottoa | Pasi Havia | Neuvottelija 416. A Finnish edition of this episode is published at www.neuvottelija.fi.

In depth

The Neuvottelija AI editions carry a long-form, fact-checked write-up of this episode: English · suomeksi.