For most of the 20th century, American corporations operated research labs where scientists created basic knowledge alongside new technologies. Since the 1980s, that role has shifted to universities and startups—a model that has been successful in software and biotechnology but has struggled in deep-tech fields that require long-term investment and large infrastructures.
America in Test Tube | Chapter 4 of 8
This article is part of the "America in Test Tube" series to mark the 250th anniversary of the United States, which examines how the world's greatest scientific power was built and the relationships between science, state, industry, and democracy.
Previous chapter: The war that built the American science funding system | for all episodes of the series
In the previous installment of the “America in Test Tube” series, we saw how World War II connected the federal government, universities, and industry and created the American research funding system. But government money and research universities are not enough by themselves to turn a scientific discovery into a product that can be manufactured, distributed, and maintained.
They have played this role for decades. Corporate research laboratoriesLarge companies did not stop at acquiring patents or improving existing products. They employed physicists, chemists, mathematicians, and engineers, allowing them to explore long-term questions and sometimes even creating basic knowledge that was not directly related to a particular product.
This system reached its peak in laboratories like Bell Labs, IBM Research, General Electric Laboratories, and DuPont. But its roots lie in the laboratory he founded Thomas Edison בMenlo Park in New Jersey in 1876 – the same year that Johns Hopkins University opened.
Edison's most important invention was the laboratory itself.
Edison's "invention factory" produced more than 400 patents, including the phonograph and a commercial electric lighting system. But economists Ashish Arora and Sharon Blenzon argue in an article inScience Because Edison's most important innovation was not a single device, but a new form of organizing the work of invention.
Rather than relying solely on the working inventor, Edison assembled a team that included mechanics, chemists, mathematicians, glassblowers, and craftsmen. The laboratory made it possible to break down a technological problem into its components, run many experiments simultaneously, and combine scientific knowledge, technical skills, and manufacturing capability.
Edison was not content with inventing products and selling the rights to them. He founded companies designed to commercialize some of his inventions and build the systems needed to operate them. The electric light bulb, for example, would have been of no practical value without power stations, transmission networks, electricity meters, switches, and lamp housings.Innovation It didn't end with the invention of the product; it required building an entire system around it.
When the inventor and the manufacturer were separate professionals
In the first half of the 19th century, there was an active market for independent inventors in the United States. Inventors developed ideas and registered patents, while manufacturing companies purchased licenses from them and turned the inventions into products.
Edison also started out this way, selling patents to telegraph companies. Railroad companies relied on inventions that came from their employees or from outside inventors. It was a relatively clear division of labor: an individual or a small workshop invented, and a company with factories and marketing power produced and sold.
But as technology became more complex and science-based, this separation became problematic. The development of synthetic dyes, materials, communication systems, electrical devices, and pharmaceuticals required knowledge that could not be acquired simply through a ready-made patent. The manufacturer was required to understand the science, continue to develop it, and adapt it to manufacturing processes.
German companies were ahead of America
The transition to organized industrial research began in part in the German chemical industry. As early as the 19s, German chemical companies recruited researchers from universities and established laboratories that were directly linked to the scientific world.
These laboratories were not content with testing a new material or improving an existing process. They created chemical knowledge on which entire families of products could be based. Society became both a consumer of science and a producer of science.
In the 20s, large American companies adopted the same principle. They realized that to operate at the forefront of technology, they had to employ scientists capable of publishing world-class research—not just engineers who tweaked a product that had already reached the market.
At General Electric, researchers like Charles Steinmetz and Irving Langmuir worked; at AT&T, physicist Clinton Davison worked; and at DuPont, chemist Wallace Carruthers developed the science that made it possible to produce new polymer materials. Scientific research and commercial development took place within the same organization.
The Golden Age of Corporate Research Labs
After World War II, the model strengthened. Federal investment made American universities world leaders, while the Cold War, the space race, and defense spending created demand for advanced technologies in electronics, communications, materials, computing, and medicine.
Corporate research labs acted as a bridge between the science generated in universities and products and infrastructure. They sometimes created the science themselves. They could maintain large, multidisciplinary teams, build expensive facilities, and wait years for research to become useful technology.
Bell Labs is the most famous example. It benefited from the unique structure of AT&T, which was a regulated monopoly in the field of telephony. The company had large resources, a stable market, and an incentive to improve the communications network over time. Under these conditions, it was possible to fund research that did not guarantee a profit in the next quarter.
The organizational structure was as important as the talent of the scientists. Long-term research requires a body that can absorb failures, maintain expertise, and connect physicists, chemists, engineers, production people, and product managers.
Why did the model begin to fall apart?
Starting in the 1980s, the architecture of American innovation changed.
The Bayer-Dol Act encouraged universities to patent and license inventions created through federally funded research. At the same time, a more sophisticated market for technology developed: companies could license, invest in, or buy a startup outright, rather than maintain their own large research lab.
The end of the Cold War also reduced government orders in areas such as lasers, communications, materials, and space systems. The public demand that helped justify long-term research investments weakened.
Another problem was knowledge leakage. A scientific paper or basic discovery by one company can also be used by its competitors. The company that bears the cost of the research is not necessarily the one that will enjoy all the profits. As knowledge becomes more mobile between companies and countries, the private incentive to fund basic research within the corporation has diminished.
Companies still wanted the “golden eggs” of innovation, write Arora and Blenzon, but became less willing to fund the goose that laid them.
The university researches, the startup translates, and the corporation buys
In place of the large corporate laboratory, a distributed system was created.
Universities produce a large part of scientific knowledge. Researchers or entrepreneurs establish startup companies To transform knowledge into technology. Funds Venture capital Fund the early stages. If the technology is successful, a large company acquires the startup, licenses it, or integrates the product into its systems.
In a sense, the economy has returned to the world in which Edison operated at the beginning of his career: one factor invents and another factor in commerce. But the similarity is only partial. Science today is much more complex, and startup companies rely on universities, advanced equipment, knowledge bases, and scientific personnel trained over years.
The new system has clear advantages. It allows companies to be quickly founded around a focused idea, to share risks among investors, and to try many approaches simultaneously. In the fields of software and life sciences, it has created a very active innovation system.
However, it is not equally suitable for every field.
Why is deep tech not like an app?
In the fields of deep tech – technologies that rely on advanced scientific research and require a long journey from the laboratory to the market – startups and venture capital are not always enough.
Developing new batteries, advanced materials, clean energy technologies, or nuclear fusion may require large test facilities, demonstration plants, complex supply chains, and investments that last for many years. Even after successful scientific proof, there is still a long way to go before reliable, cheap, and industrial-scale production.
A startup can prove that an idea works in the lab, but it can't necessarily establish an entire industry on its own. Investors may not want to wait ten or twenty years, and a large company that has already given up its internal research capability may have difficulty even evaluating the technology being offered.
This is a paradox: corporations withdrew from research because they could buy innovation from outside, but after years of withdrawal they may lose the knowledge needed to identify which innovation is worth purchasing and implementing.
When a large company maintains internal scientific and engineering capability, the distributed system can work well. The university generates knowledge, the startup develops a prototype, and the corporation brings the technology to production and the market. When internal capability disappears, the bridge between science and commerce weakens.
You can't just rebuild Bell Labs.
The solution is not necessarily to recreate the 1960s. Bell Labs emerged from unique historical conditions: a regulated monopoly, stable revenues, large government demand, and a national technological system that took decades to develop.
It is impossible to order a modern society to establish such a laboratory and hope that the same conditions will return.
The authors suggest focusing on areas where the bridge between science and commerce is weak and trying a variety of institutions: focused research organizations, Advanced Research Agency-style programs, public translational research institutes, and hybrid government-industry partnerships.
A focused research organization can bring together a multidisciplinary team around a problem that is not suitable for a regular academic grant but is also not ripe for a startup. A translation institute can take a promising discovery and move it through the engineering, testing, and demonstration stages. Public funding can bear some of the risk where the private market has difficulty waiting.
There is no single solution that fits every industry. The goal is not to choose between universities, startups, and corporations, but to strengthen the connections between them.
Science is not a product that can be ordered at the last minute.
The story of corporate research laboratories teaches that innovation is not an automatic chain. A scientific discovery does not automatically become a product, and a successful product is not born solely from the brilliance of a single entrepreneur.
Institutions are needed that are capable of preserving knowledge, connecting fields, building infrastructure, and bearing risk over time. Sometimes it is a university, sometimes a corporate laboratory, sometimes a young company, and sometimes a public body. In most cases, a combination of them is required.
Research universities remain the foundation on which the system rests, but not every task can end with the publication of an article or the registration of a patent. For knowledge to become technology, translation capabilities are required: engineering, manufacturing, testing, standardization, and understanding the market.
Edison’s Invention Factory was an early attempt to bring all these steps together under one roof. Today’s system has scattered them among many institutions. The challenge for science and industrial policy is to ensure that the links between them are strong enough – and that there is no gap left between the laboratory and the factory that no one is willing to finance.
For the scientific article: Opening the scientific article
More on the subject on the science website
- America in Test Tube – Chapter 1 – 250 Years of Promise: American Science Born as Part of the Democratic Idea
- 250 years of the United States: This is how the world's greatest scientific power was built — and how it endangers its future
- The AI Arms Race: Why Anthropic Wants to Hit the Brakes
- Hungarian researchers warn of long-term impact of disconnection from EU research programs under Orbán
- Artificial Intelligence in Israel Between Local Challenges and Global Trends: Neeman Index Dashboard Presents Continuous Measurement
