Siboniso Nxumalo | Chief Investment Officer | Old Mutual Investment Group | mail me |
Electricity transformed the world, not simply because it existed, but because of what it made possible. That distinction may prove important as we think about Artificial Intelligence (AI).
In 1882, Thomas Edison built the world’s first commercial power station at 255 Pearl Street in Lower Manhattan, New York. The incandescent light bulb was revolutionary. However, the real breakthrough was the system behind it: the generators, cables, meters, and distribution network. These components moved electricity from laboratories into homes and businesses.
By 1900, investors knew electricity would change the world. However, they did not know which companies would win, how long it would take, or where profits would accumulate.
The economic winners were not always obvious. Electricity utilities became essential, but many of the greatest profit pools emerged in factories, new industries, and business models that electricity enabled. The same applies to AI. We know it will change the world, but we do not know where its enduring economic benefits will settle. The principle that technology changes the world is clear. However, identifying where lasting value emerges remains far more difficult.
Today, semiconductor designers, equipment manufacturers and infrastructure providers are earning extraordinary returns from the AI buildout. However, investors should not assume that tomorrow’s profit pools will resemble today’s. Many companies at the centre of the AI narrative remain only modestly profitable or loss-making. Consequently, their valuations depend heavily on assumptions about future cash flows.
Every era has its transformational technology
In the 1920s, radio was the internet of its time. Its adoption was explosive. US radio stations grew from just five in 1921 to more than 500 by 1923. Meanwhile, equipment sales rose from $60 million in 1922 to more than $840 million by 1929. Radio had become one of the defining technologies of its age.
RCA stood at the centre of this transformation. Between 1921 and its 1929 peak, RCA’s share price rose almost 100-fold as investors bet that radio would change the world. They were right about the technology’s potential. However, by 1932, RCA’s share price had fallen almost 98%, despite radio’s continued rapid spread.
The lesson is that technological success and investment success are not always the same. Investors may correctly identify the future and still misjudge where the economics will settle. In other words, even when technology changes the world, it does not guarantee that investors will capture the resulting value.
Yes, AI is here, but something unusual is happening
The world’s largest software companies have started behaving more like infrastructure businesses. Microsoft, Amazon, Alphabet, and Meta are undertaking one of the largest private infrastructure buildouts in modern history. They are spending heavily on data centres, semiconductors, servers, networking equipment, land, power and cooling.
The numbers are striking. Microsoft’s net property and equipment rose from $135.6 billion in 2024 to $205 billion in 2025. Much of this increase relates to data centres.
These companies are reinvesting more of the cash they generate into physical infrastructure. Consequently, their balance sheets, once dominated by intangible assets, increasingly contain tangible infrastructure and land. Their language remains digital, but their economics are becoming industrial.
For decades, winners in the digital economy became more capital-light as they grew. Today, AI leaders appear to be becoming more capital-intensive. This raises a profound question for investors: what happens to returns on capital, profitability, and valuation multiples when software businesses begin to resemble infrastructure businesses?
AI is often described as the next software revolution. However, the scale of investment suggests another possibility. AI may prove less like the internet and more like electrification.
Electricity became a foundational input across industries worldwide. Factories did not become electricity companies; they became factories powered by electricity. Similarly, banks, retailers, asset managers and manufacturers may become businesses powered by AI.
If so, the largest profit pools may not accrue exclusively to those building AI. Instead, they may emerge in industries and business models that use AI to reshape how they create value. This is another reason to consider how technology changes the world beyond the companies that develop it.
People tend to imagine the future as an improved version of the present. They picture a faster horse, an idea often attributed to Henry Ford, who reportedly said, “If I had asked people what they wanted, they would have said faster horses.”
Yet major technological shifts can overturn the assumptions on which entire industries operate. We are already experiencing AI’s benefits as an asset manager. We can process more information, analyse more companies, and investigate opportunities across more geographies with the same number of analysts. However, this may still represent the “faster horse” stage of AI. At this stage, businesses use new technology to perform existing tasks more efficiently.
The larger opportunities may emerge from businesses that use AI to change how an industry operates, how it creates value, and where its profit pools accumulate. The real investment opportunity may therefore lie in understanding how technology changes the world by transforming entire business models, rather than simply improving existing processes.
History teaches us that profit pools do not simply grow; they move
The internet did not merely expand existing profit pools. It disrupted them, redirected value away from established players, or created entirely new ones. AI may do the same.
As investors, our task is to monitor where AI disrupts, enhances, or reprices profit pools. We cannot confidently predict where the greatest value will accrue. However, AI could materially change the economics of many industries.
Some of the most important beneficiaries may not be today’s AI leaders, including model developers, semiconductor companies, and cloud providers. They may not even exist yet.
AI may become as foundational as electricity. If it does, the largest winners may be those that use it to rewrite the rules of the game. Technology changes the world, but investors must look beyond the technology itself to understand who benefits from that transformation.
AI may be inevitable. Investment returns are not
Pearl Street in New York offers a useful reminder. Investors recognised electricity’s transformative potential, but they could not know where its economics would ultimately settle.
AI may follow a similar path. The technology appears transformative, adoption seems inevitable, and the investment opportunities are enormous. Our task is not simply to predict whether AI will change the world. Instead, we must identify where profit pools migrate, who captures them, and whether today’s prices reflect those possibilities.
History rarely provides answers. However, it often provides better questions. The most important question today is not who is building AI, but who will ultimately use it to change the rules of the game. Technology changes the world. Economics determines the winners.
