What the history of general-purpose technologies suggests about who benefits from the AI build-out
Artificial intelligence has transformed financial markets before transforming the broader economy. The two events are related, but they are not the same. Over the past three years, the equity value of firms most closely associated with the AI opportunity, from platform companies to semiconductor leaders and infrastructure enablers, has increased at a pace and scale that has few close analogues in the recent history of capital markets. A handful of technology companies now account for a share of global equity market capitalization that would have seemed implausible only five years ago. Private capital markets have moved in parallel, with valuations of specialist model developers, infrastructure operators, and enabling suppliers reaching levels that reflect assumptions about future economic contribution rather than current earnings.
The transformation of the real economy will unfold over a longer horizon than the transformation of capital markets, as it has done in every prior technological wave for which comparable evidence exists. For institutional investors, the fact of artificial intelligence changing the global economy is now the consensus view. The more consequential question concerns how the returns from that change will accrue over the decade ahead, and the distribution is likely to keep moving. It has done so in every previous wave of general-purpose technology adoption, from the railways to electricity to the internet, and there is little in the current wave that suggests a departure from that pattern.
A Concentration of Value in the First Phase
The most visible feature of the current period is the concentration of value among a small number of technology firms. The small group of public companies most identified with the AI build-out now represents a share of global equity market capitalization that reflects both their current earnings power and the market’s assessment of their positioning for the next phase of the technology’s diffusion. Their combined market value is comparable in aggregate to the equity value of many national capital markets. This concentration is unusual, though not unprecedented; similar concentrations have appeared in the early phases of prior technological transitions, supported in each case by the market’s judgment that the leading firms possessed durable positions in the emerging technology.
Public market re-rating has been reinforced by private market activity. Specialist model developers have raised substantial rounds at valuations that were until recently reserved for mature technology companies. Infrastructure operators have accessed public and private capital at scale. The past two years have seen renewed public and private market activity among specialist operators, chip designers, and cloud infrastructure providers, with further offerings anticipated or announced. Each of these transactions reflects a market judgment that the position being financed will be durable enough to justify long-duration equity capital, on terms that already assume substantial future economic contribution.
What the concentration of value has produced, in aggregate, is a first-phase distribution of returns heavily weighted toward the firms developing and deploying the technology at the leading edge. This is consistent with historical experience of general-purpose technologies at similar stages, and it may persist for some time. It is also unlikely to remain the final distribution. Prior waves have seen substantial redistribution of value across subsequent decades, and the pattern of that redistribution has particular relevance to how institutional capital positions itself for what follows.
The Lessons of Prior Technological Waves
The economic history of general-purpose technologies is unusually helpful in framing what is likely to follow. Four waves stand out as particularly instructive. The railways in the second half of the nineteenth century transformed the movement of people, goods, and information across major economies. Electricity transformed manufacturing, urbanization, and household life over the first half of the twentieth. The internet transformed communication, commerce, and information distribution across the last three decades. Each wave produced initial concentrations of value among the developers and immediate operators of the technology, followed by extended periods during which value diffused across a much wider set of industries and geographies.
In the railway era, early value concentrated in the operators of trunk lines, in the equipment manufacturers, and in the financial institutions that intermediated the enormous flows of capital required to build the network. The subsequent decades saw much of the economic gain accrue to the industries and cities that the railways enabled, from agriculture to manufacturing to retail commerce. In the electrification era, generation and equipment companies captured the first phase, while the industries that reorganized around cheap and reliable power captured the second and more distributed phase. In the internet era, infrastructure providers, dominant platforms, and semiconductor firms captured the first two decades of value, while the diffusion into productivity gains across the wider economy has been slower, more uneven, and still ongoing.
Two features of these prior waves are worth highlighting. The first is that the infrastructure supporting a general-purpose technology can capture a substantial share of the value, sometimes exceeding that of the operators of the technology itself. The classic pattern of railway equipment suppliers outperforming several of the railway operators has recurred in later waves. The second is that capital suppliers, including the financial institutions that intermediated the required investment, have frequently done well across all phases, particularly where they participated in both the infrastructure and the enabled industries. Neither pattern was accidental. Both reflect the specific economics of assembling and financing very large stocks of long-duration physical capital under conditions of rapid technological change.
The Slow Diffusion into the Productive Economy
The mechanism by which a general-purpose technology transforms the wider economy has been studied carefully across prior waves, and it is more consistent than the specific applications suggest. Initial adoption is uneven, concentrated in industries and geographies with the capacity and incentive to invest. The productivity gains associated with the technology require complementary investments in skills, business processes, and supporting infrastructure that most organizations acquire slowly. Employment and output effects follow, with significant transitional friction that varies by industry and by workforce. Aggregate productivity growth tends to lag the technology’s arrival, in some cases by a decade or more, before it becomes visible in national statistics.
The current wave shows signs of following this pattern. Adoption has been most rapid in sectors with a strong combination of digital data, established analytical processes, and clear commercial application, including software development, financial services, professional services, and specific segments of manufacturing. Consumer adoption has been faster than in previous waves, largely because the interface presented by the current generation of models is more accessible than the interfaces of earlier technologies. Broader economic effects, including sustained productivity gains, employment shifts, and reorganization of specific industries, are likely to unfold over a longer horizon than the current pace of financial market movement suggests.
The Three Layers of Return
Three distinct layers of return are emerging within the current wave, and their relative performance over the next decade is unlikely to resemble the pattern established over the past three years. The first layer consists of the AI companies themselves, including the model developers, application specialists, and the largest platform companies whose services depend directly on artificial intelligence. This layer has captured the largest share of value in the first phase. Its risk-return profile is characterized by rapid product cycles, meaningful competitive erosion between successive model generations, and a wide dispersion of eventual outcomes among competing firms.
The second layer consists of the infrastructure providers that support the technology, including the operators of data centers, the specialist chip designers and manufacturers, the power generation and transmission companies that supply the electricity required by the workloads, and the network operators that connect them. Historical experience suggests that this layer often captures a share of value larger than what its current market presence implies. Infrastructure has been the beneficiary of every prior technological wave in which the underlying physical capital was large and durable. The AI build-out sits within that pattern, with the additional feature that the required capital stock is unusually specialized and capital-intensive.
The third layer consists of the capital suppliers, meaning the institutional investors, banks, insurance companies, and sovereign vehicles whose deployment of long-duration capital is required to finance the build-out. Capital suppliers benefit through the yield and equity returns from the infrastructure they fund, the intermediation fees from arranging complex cross-border transactions, and the participation rights they secure through co-investment structures. Their relative position tends to be less volatile than that of either the technology developers or the infrastructure operators, because they receive returns from a diversified stake in the entire build-out rather than from a specific commercial position. In prior waves, disciplined capital suppliers have often been among the most durable beneficiaries.
Implications for Institutional Capital
For institutional investors, the practical implication of the layered structure is that the distribution of returns is likely to remain in motion over an extended period. Portfolio construction that overweights the current concentration among the leading technology firms may miss the redistribution that historical experience suggests will follow. Underweighting overall exposure to the technology carries a different risk, given the pace at which artificial intelligence is being embedded in the productive economy. The appropriate frame includes all three layers, with attention to which layer is currently priced accurately relative to its historical role.
The frame also has implications for time horizon. The concentration of value at the technology developer layer has been the story of the past three years, and it may continue for some time longer, but the redistribution phase in prior waves has typically played out over one to three decades. Institutions with the mandate and the balance sheet to operate at that length can construct positions that participate in each phase of the distribution. Those with shorter horizons face a more difficult task, since the specific timing of the redistribution is less predictable than its general direction.
The Question of Who Benefits
The question of who benefits from artificial intelligence has more than one answer, and the answers are likely to change as the wave unfolds. The first phase has been dominated by a small number of technology firms whose equity value has re-rated dramatically. The next phase is likely to see more value accrue to the infrastructure that supports the technology, to the capital suppliers who finance it, and to the industries and workers able to adopt it productively. The concentration observed today is unusual by historical standards, though the historical standards themselves suggest it is unlikely to be permanent.
The transformation of the global economy by artificial intelligence has begun in capital markets, as such transformations typically do, and its unfolding across the productive economy will take longer. For institutions positioning themselves within this wave, the practical work involves constructing an approach that recognizes its layered structure and that is prepared for the redistribution that has followed every prior wave of comparable scale. The technology developer story will continue to attract most public attention, and it will remain important. The infrastructure story is entering its most capital-intensive phase, and its economics are becoming more clearly defined. The capital supplier story is quieter, but has often produced the most durable returns across prior waves. The historical record is clear about the general pattern of the distribution, and institutions that build their frameworks around it will be positioned to participate through each of its phases.
About Berkeley Financial
Berkeley Financial is an international financial group providing institutional banking, private banking, custody, and cross-border financial solutions. With a focus on governance, relationship-driven execution, and multi-jurisdiction expertise, Berkeley supports institutions and sophisticated clients with international financial needs across key markets, including Latin America, Europe, and the United States.
Disclaimer
This article is provided for informational purposes only and does not constitute investment, legal, tax, regulatory, or financial advice, nor an offer, solicitation, or recommendation to buy or sell any security, financial instrument, investment product, or infrastructure asset. References to sectors, financing structures, and market trends are general in nature and may change over time. Institutions should evaluate any investment, financing, or strategic decision based on their specific objectives, risk tolerance, jurisdiction, and applicable regulatory requirements.
