History says AI is likely to widen the gap between corporate giants and everyone else: Goldman Sachs

History suggests the boom in artificial intelligence could strengthen the competitive edge of America's biggest and most dominant companies, according to Goldman Sachs.

As the world grapples with how AI could reshape the global economy, a major question is whether it will disrupt the advantages of today’s biggest winners or deepen them.

Goldman’s view — based on nearly a century of income, sales, and corporate tax data as well as other records — leans toward the latter.

“Corporate concentration in the US has steadily climbed since the 1930s, rising more rapidly during periods of faster technological change,” Goldman Sachs chief economist Jan Hatzius and his team wrote in a new report.

“The historical lesson from previous technology shocks is that new technologies and the successful investment in intangible capital needed to deploy them have tended to raise concentration as scale and network effects accrue to leading firms,” the team added.

The finding complicates some of the worries that investors have grappled with so far this year.

A February report from independent research firm Citrini went viral, briefly dragging major tech and financial stocks lower. It laid out a bleak vision for how the AI boom could lead to mass disintermediation, triggering white-collar layoffs and ultimately a sharp stock market downturn.

While acknowledging AI could increase competition, the Goldman Sachs report, by contrast, examines the opposite possibility.

Over nearly a century, sales and profit margins have increasingly accrued to the biggest US companies. And during periods of rapid tech advancement, companies with bigger scales have most often had the means to ride the technological waves.

“New technologies tend to involve high fixed deployment costs and low marginal costs of scaling, so that firms with the capital and organizational capacity to make the upfront investments required to adopt them — including in data infrastructure, software, and organizational redesign — can spread those costs across a larger output base, gaining market shares over smaller rivals,” Goldman Sachs said in the report.

Read more: How to protect your portfolio from an AI bubble

For years, a near-constant task for AI watchers has been unpacking who's likely to win and lose. Leading AI model makers Anthropic (ANTH.PVT) and OpenAI (OPAI.PVT) are rushing to potential IPOs later this year in a race to gain more capital to finance their computing needs.

Meanwhile, a handful of major tech firms are also planning to spend more than $700 billion this year and over $1 trillion before the decade’s end to build out AI infrastructure.