How Industrial Research Labs Turned the U.S. Into a Tech Superpower - and Who Got Left Behind

The standard narrative of American technological leadership points to a long parade of inventions: the telephone, the light bulb, the transistor, the internet. But a detailed analysis of 1.6 million patents spanning nearly 150 years suggests the story is more structural than that. The United States did not gradually accumulate its way to technological dominance. The shift happened abruptly, over a compressed period in the early 1920s, and was driven primarily by how innovation was organized rather than by any particular discovery.

The study, published in Research Policy and authored by Frank Neffke from the Complexity Science Hub (CSH) alongside colleagues from the Growth Lab at Harvard University, involved digitizing hundreds of thousands of pages of historical documents covering patents and inventors from 1856 to 2000. The scale of the data effort allowed the researchers to move from anecdote to systematic pattern.

What changed in the early 1920s

For most of the 19th century, American innovation was a craft-based activity. Individual inventor-entrepreneurs - working often with small circles of trusted collaborators, sometimes connected through family networks - drove technological progress through a combination of practical skill, experimentation, and accumulated experience. The system produced important advances. It also had a ceiling.

The industrial research lab, an organizational form that originated in the German-speaking world, crossed the Atlantic in earnest after World War I. American firms began hiring teams of specialized engineers and scientists who worked together systematically, drawing on formal scientific knowledge rather than primarily on practical know-how. The shift from craftsman to engineer as the central figure of innovation happened fast. By 1945, engineers made up just 0.7 percent of the U.S. population but accounted for 25 percent of all patents.

"We analyze systemic shifts in the way invention was organized in the US, supported by a massive data effort," Neffke said. "At the center of this transformation was an organizational innovation: the industrial research lab - an idea born in the German-speaking world that rapidly diffused in the U.S. after World War I."

Teamwork as the new engine

Research labs reorganized how people worked together on invention. Teams in labs were more likely than independent inventors to collaborate repeatedly over time, to work across long distances, and to produce what economists call "neue Kombinationen" - novel combinations of existing technologies that create something genuinely new. The lab format enabled a more systematic approach to exploring the space of possible innovations.

Invention also became more explicitly science-based. Laboratory workers drew on formal physics, chemistry, and engineering principles rather than relying primarily on craft knowledge. This represented a decisive change in the epistemology of American innovation - from knowing how to do things to understanding why they worked, and then using that understanding to do new things.

Where the labs concentrated activity

Industrial research labs also reshaped the geography of innovation. During the late 19th century, inventive activity had spread geographically, moving from large cities into smaller towns as transportation and communication improved. The emergence of research labs reversed this trend, reconcentrating innovation in a small number of large metropolitan centers that could support the specialized labor markets and institutional infrastructure that labs required.

"This shift helped fuel the rise of a small number of large cities in what we now know as the American Rust Belt, but which in its heyday was the Silicon Valley of the early 20th century," Neffke noted.

A system that excluded as it advanced

The new innovation system was not equally open. The shift to professionalized, science-based invention in corporate labs created barriers that had not existed under the craftsman system. Women and foreign-born inventors, who had participated meaningfully in the earlier craft-based model, became significantly underrepresented in the emerging laboratory environment. Professional credentials, academic networks, and hiring practices in the new labs favored a narrower demographic.

These exclusions were not incidental. They were built into the structure of the new system and persisted for decades, suggesting that changes in how innovation is organized have social consequences that extend well beyond their technical effects.

Labs declined, then returned

Industrial research labs did not maintain their dominance indefinitely. The researchers found that after the 1970s, firm-based teams within labs underperformed independent teams in producing genuinely novel combinations. Labs became better at refining and extending existing technological trajectories than at opening new ones.

The current moment shows familiar patterns. Tech giants including Google, Meta, and Amazon have rebuilt large-scale research operations that in some respects mirror the Bell Labs model of the mid-20th century. AI capabilities are emerging primarily from these corporate labs rather than from academic or independent inventors. Whether this revival will reproduce the exclusionary dynamics of its predecessor - and whether it will similarly reshape economic geography - are questions the historical analysis raises without fully answering.

"We often describe the history of technology as a succession of technological breakthroughs," Neffke observed. "However, our study suggests that social innovations may be just as important."