Analysts at Goldman Sachs say the net effect of artificial intelligence on competition and corporate margins will depend on how advantages distribute across firms. In a research note, they frame two broad paths: one where dominant incumbents strengthen their positions, and another where the most adept adopters widen the gap by outperforming rivals.
The note highlights three broad forces that are likely to determine AI's impact on concentration and profitability: the extent of globalisation, the prevailing regulatory environment, and the role of economies of scale. These elements will interact with how AI technologies are deployed within industries, and with the nature of the industries they disrupt.
Goldman Sachs draws parallels between the economics of the AI era and earlier technology-driven revolutions such as electricity and the internet. In those episodes, the economy experienced substantial productivity improvements overall, but the distribution of gains was uneven. Infrastructure owners and firms controlling critical platforms tended to capture a disproportionate share of profits.
However, the analysts caution that AI could follow a different path if the relative importance of raw model quality declines over time. They note that if frontier AI models become widely commoditised, open-source efforts and model diffusion could compress margins that are attributable purely to model performance. In that scenario, competitive advantage would increasingly hinge on non-model assets.
Those non-model assets include proprietary enterprise data, tight integration into customer workflows, established distribution channels, and customer lock-in. Firms that combine these elements may be better positioned to secure persistent profitability even as core model technology becomes more accessible.
Long-run trends in corporate concentration across advanced economies are presented in the note as background to assessing AI's potential effects. Goldman Sachs points to more than a century of tax and administrative records that show rising concentration, particularly in the United States. In the U.S., the expansion of larger firms in finance, manufacturing, information, retail, and later wholesale trade has driven a steady increase in concentration since the 1930s.
To explain this rise in concentration, the analysts set out three competing hypotheses. One argument credits globalisation and expanded trade for enabling large firms to capture a disproportionate share of growth opportunities. A second contends that weaker antitrust enforcement and higher regulatory barriers have insulated incumbents from competition. A third hypothesis asserts that economies of scale allow a subset of firms to grow larger and seize market share.
Goldman Sachs finds the economies-of-scale explanation the most persuasive. The note emphasizes that concentration typically accelerates during periods of rapid technological change. In such episodes, productivity dispersion widens as frontier firms extend their lead, especially in sectors that produce technology and in industries that are deeply transformed by new technologies, such as online retail.
Despite rising profit margins over recent decades, the analysts estimate that only about one-third of that increase can be explained by greater concentration. They point to higher markups driven in part by rising incomes that make consumers less price-sensitive and by diminished price-comparison activity across sellers. Rising concentration has reduced competition at the national level, even if local competitive dynamics differ.
When applying these observations to AI, Goldman Sachs underscores the technology's two-sided implications. Industries currently most exposed to AI tend to be more concentrated and to enjoy higher margins. AI-driven disruption may intensify competition within those sectors as firms deploy new capabilities against incumbents. At the same time, the introduction of new technologies and the intangible capital required to deploy them often reinforce scale and network effects, enabling leading firms to consolidate advantages.
Takeaway
The ultimate direction of AI's impact on concentration and corporate profitability is not predetermined. Outcomes will reflect how model technology evolves, whether frontier models remain differentiated or become commoditised, and which firms can leverage data, integration, distribution, and customer relationships to convert AI capabilities into durable profits.