The Influence of Bell Labs

We’ve talked previously about Bell Labs’ long, storied history as an innovation engine and a generator of new technology. For decades, it spun off new major inventions and scientific discoveries as part of its mandate to help build AT&T’s telephone network.

Bell Labs was notable for combining the best aspects of academic and industrial research. As in a conventional academic lab, scientists had the freedom to pursue research avenues they found promising without being bound by concerns of immediate profitability or return on investment. It was understood that efforts might take years to pay off, and only a handful would. Researchers could also publish papers and engage with the broader intellectual community in their field, while using cutting-edge equipment that only an industrial lab could provide. And they could pursue their research without having to worry about teaching classes or applying for grants.

Discussion of Bell Labs tends to center around the impact and successes of the Labs itself. But a perhaps underappreciated impact of Bell Labs is the influence it had on other large corporations. Bell Labs was highly prestigious, and its invention of the transistor demonstrated that world-changing products could come by funding “basic” scientific research. Inspired by Bell Labs, in the second half of the 20th century, a variety of corporations started their own research operations based on the Bell Labs model.

Bell Labs and Industrial Research

Industrial research labs started to become popular in the US around the turn of the 20th century. Along with Bell Labs, companies like DuPont, General Electric, and Kodak all employed scientists in research labs, and by 1940 there were over 1,000 industrial research labs in the U.S. Most of these labs were small (on average they employed fewer than 100 people), and as late as 1940 there were fewer than 30,000 people employed in a “scientific” capacity in all US manufacturing.

Research labs in the U.S. over time.

But following WWII, research spending by both the government and the private sector increased enormously. In 1946, private R&D spending in the U.S. was about $500 million annually, roughly what it had been before the war. By 1951 it had increased four-fold to $2 billion annually, and between 1953 and 1977, in real terms, it quadrupled again.

Unpacking the reasons for this increase goes beyond the extent of this essay, and it certainly wasn’t just because of AT&T and Bell Labs. But Bell Labs’ achievements, particularly the invention of the transistor in 1947, seem to have prompted the decisions of many companies to start their own R&D labs, and influenced how those labs were structured.

There seem to have been several modes of influence. For one, the transistor itself ushered in a new world of semiconductor devices. Semiconductors were a product of deep, scientific understanding of the physical nature of matter. Anyone who wanted to compete in the new market by developing their own semiconductor-based products would need to acquire the relevant scientific expertise.

The transistor also demonstrated that “basic” research could result in enormously successful, world-changing products. The world had just seen the amazing power of scientific research in wartime achievements like radar and the atomic bomb, and the transistor showed that the fruits of such research weren’t limited to enormous government projects. The success of nylon, another successful product that was the result of basic research by DuPont, also reinforced this perception.

Finally, Bell Labs was incredibly prestigious, and companies wanted to burnish their reputations by having their own scientific research establishments. Not only was this prestige desirable in its own right, but it could allow companies to acquire the sort of talent that they otherwise might not be able to. The transistor cemented Bell Labs’ reputation as one of the best research labs in the world, which allowed it to acquire some of the world’s best scientific talent, which resulted in further scientific achievements.

Its prestige, reputation for excellence, and envious working environment allowed Bell Labs to acquire some of the most talented researchers in the world. Bell Labs Nobel Prize winner Horst Stormer noted that “Over a very long stretch of time, it was the best place in the world and it attracted — and attracts — the best people.” In his short memoir about Bell Labs, Michael Noll likewise noted that “it seemed everybody wanted to get a job there.” Of Bell Labs’ 10 Nobel Prizes, 8 came from researchers hired in the ‘50s, ‘60s, and ‘70s following the invention of the transistor.

Bell Labs thus created a new world that required some degree of scientific capability, showed what could result from creating such capabilities, and provided a playbook for how to do it. In the 1950s and beyond, many companies started their own research operations on the Bell Labs model: creating an academic-esque environment and giving researchers freedom to follow their interests, without necessarily needing to worry about marketable products or immediate profitability.

Several of these research labs, unsurprisingly, were founded by companies in the computers and electronics industry. IBM, for instance, had a research division prior to the 1950s, but it was mostly a small group of electrical engineers who were focused on figuring out how to build new products. Thomas Watson Jr., son of the founder of IBM and its former CEO, describes it in his autobiography:

IBM’s main laboratory, on North Street in Endicott, was a very peculiar place. Between three hundred and four hundred people worked there, but the whole thing was built around seven senior engineers whom Dad called his “inventors”... When [Dad] had an idea for a product, he’d call in one or two of these old birds and describe what he wanted it to do. Then the inventors would go back and try to “put it in metal” as they used to say.

At the time, IBM primarily produced mechanical punchcard-based calculators, but Watson Jr, recognizing that the future was in electronics, scaled up their research operations and hired thousands of scientists and engineers who understood semiconductors and solid-state physics. To organize a program of “pure research," IBM hired Emanuel Piore, the former head of the Office of Naval Research:

Piore gave a jolt to some of our product development engineers as well. They were like sprinters encountering their first marathon runner, and were amazed to see IBM start funding experiments in exotic fields that seemed unlikely to bear fruit for decades, if ever — like superconductors and artificial intelligence. What the engineers thought of as basic research Piore often dismissed as mere long-term product development, and what he called research was so far removed from what the engineers were doing that they saw no reason for it at all. At Piore’s urging we doubled the percentage of our revenues devoted to research and development, and much of the additional spending was earmarked for pure science.

IBM Research went on to spawn a variety of major scientific and technological discoveries. From Wikipedia:

IBM Research's numerous contributions to physical and computer sciences include the Scanning Tunneling Microscope and high-temperature superconductivity, both of which were awarded the Nobel Prize. IBM Research was behind the inventions of the SABRE travel reservation system, the technology of laser eye surgery, magnetic storage, the relational database, UPC barcodes and Watson, the question-answering computing system that won a match against human champions on the Jeopardy! television quiz show.

When setting up its research efforts, IBM consciously took notes from Bell Labs. An official company history of IBM’s history of innovation notes that “IBM had a model: AT&T’s Bell Laboratories”. Another book published by IBM about its research commercialization strategies likewise notes that “Originally, the IBM Research division operated as a separate entity, patterned after Bell Labs.”

Another electronics manufacturer that patterned its research efforts after Bell Labs was Texas Instruments. Gordon Teal, who invented single-crystal pulling at Bell Labs, was hired by Texas Instruments in 1953 to run its newly established research labs, the Central Research Laboratory. An official company history states that “The strength of CRL was influenced by the basic research model that Teal had learned at his previous job with AT&T's Bell Labs.” In an oral history, Teal describes how researchers were allowed to pursue their own projects that they felt were promising:

Goldstein: You were saying that by the late 'fifties and early 'sixties, you were spending less time in the lab. Where did the direction for the laboratory come from? Who was in charge of the projects that they were working on?

Teal: No one. As the people that I brought in got experience, they were capable of having good ideas themselves and not have to be directed in everything they did.

Goldstein: They would launch their own research projects?

Teal: Yes, and I depended on the group heads to give them instructions rather than me.

While these efforts are described as “basic research," they don’t appear to be quite as sequestered from product development as in some other Bell Labs-inspired research labs. In the oral history, Teal notes that the labs’ efforts weren’t quite pure science, and research directions tended to be product-focused. They nevertheless appear to be highly influenced by Teals’ Bell Labs experience. Research at Texas Instruments would ultimately result in the first production silicon transistor, and the co-invention of the integrated circuit.

Perhaps the most famous research lab inspired by Bell Labs was Xerox’s Palo Alto Research Center (PARC). Concerned that their copier business would potentially be undermined by advancing computer technology, Xerox acquired computer manufacturer SDS in 1969 to help it stay at the forefront of modern technology. Following the purchase, chief scientist Jack Goldman convinced company leadership to build a new research operation. From Dealers of Lightning, a history of Xerox PARC:

On the surface the rationale for the so-called “Xerox Advanced Scientific and Systems Laboratory” was to fortify the new subsidiary’s weak research capability. But from that foundation Goldman was intent on building a much larger edifice. Cannily recognizing that Xerox yearned to be ranked alongside such paragons of industrial muscle as IBM and AT&T, he sketched out a corporate research center engaged in basic science independent of any existing product group, exactly like IBM’s fabled Yorktown Heights research center and AT&T’s Bell Laboratories. About half the staff would be devoted to advanced physics and materials research, and the rest to the new sciences of systems and computing.

PARC would be founded “by men whose experience had taught them that the only way to get the best research was to hire the best researchers they could find and leave them unburdened by directives, instructions, or deadlines.” 

For the most part, the computer engineers of PARC were exempt from corporate imperatives to improve Xerox’s existing products. They had a different charge: to lead the company into new and uncharted territory.

PARC would become famous as the birthplace of much of the technology behind the PC revolution: out of PARC came the first personal computer, the graphical user interface, ethernet, and the laser printer.

But not every research lab inspired by Bell Labs was at a computer or electronics manufacturer. PARC was established by Jack Goldberg, who had formerly led the science arm of Ford’s Scientific Research Lab, which had been founded in 1951 and was intended “to be a major scientific institution producing innovation," pursuing basic research independent of any product development. Ford’s research efforts also seem to have been inspired by Bell Labs:

…The lab operated according to a philosophy that was established by AT&T’s Bell Labs and IBM’s Thomas J. Watson Research Center and that has now been essentially abandoned. These great institutions pursued research topics not because they were likely to contribute to the parent company’s bottom line anytime soon but because the corporation believed that research for research’s sake was something a real company did.

“We had the freedom to do what we were interested in,” says Arnold Silver, who worked in the Ford lab in its heyday. “We could follow our noses — and particularly, we could follow the data.”

Among the achievements of Ford’s research labs are the SQUID (an extremely sensitive magnetic field sensor) and the sodium-sulfur battery.

Another Bell Labs-inspired research lab outside the electronics industry was Exxon’s. In the early 1970s Exxon was worried that petroleum would begin to be exhausted in the coming decades, and was willing to fund long-term research in the hunt for alternative energy technologies. These efforts were wide-ranging, including “fuel cells, solar cells, computer chips, superconductors, and batteries, as well some non-energy projects, such as fax machines and word processors.” Exxon researchers had a great deal of freedom, access to the best equipment, and weren’t burdened with questions of profitability or return on investment. From The Long Hard Road, a history of the lithium ion battery and the electric car:

…Life at Exxon Research and Engineering was turning out to be a lot like life at Stanford. Exxon Research featured a highly academic atmosphere — groups of PhDs in small labs, surrounded by more small labs with more PhDs, mostly doing chemical and solids research. Moreover, Exxon’s lab was legally considered a not-for-profit entity, so there was no pressure to produce any sort of short-term economic benefit.

Exxon felt that it was in direct competition with Bell Labs, and acted accordingly:

Exxon corporate management felt they were in a head-to-head competition with Bell Labs, which was about twenty miles down the road in Murray Hill, New Jersey. Edward E. David, who was later named president of Exxon Research and Engineering, actually saw it as a point of pride. David had a doctorate from MIT in electrical engineering. He had previously served as a scientific advisor to President Richard Nixon and had spent twenty years at Bell Labs. And he wanted Exxon Research to be better than Bell Labs. Better, in fact, than any corporate lab in the world. Moreover, he kept score, not by counting dollars but by counting scientific papers and patents.

Exxon research would later become infamous for predicting the climate impacts of greenhouse emissions, then suppressing the findings, but it also created the first rechargeable lithium ion battery.

Conclusion

Bell Labs ultimately wasn’t able to maintain its long-horizon research environment following the dissolution of the AT&T monopoly, and it seems like the same is true for the Bell Labs-inspired research organizations. IBM still funds a great deal of research, but is much more tightly integrated with product development, and its major science and technology awards mostly seem to be for work done decades ago. Likewise, Exxon spends more than a billion dollars annually on research, but via a more carefully managed “stage-gating” process, where “Researchers partner with the business lines to determine the business benefit of a technology, establish research and development goals and timelines, steward independent project reviews and authorize project funding.” In Dealers of Lightning, Michael Hiltzik argues that by the 1990s PARC was no longer engaged in such unrestricted research decoupled from product development.

Evolution of IBM’s research process, from Innovation Passport.

This probably shouldn’t be surprising. I argued in my earlier piece on Bell Labs that only a fairly unique set of historical circumstances allowed Bell Labs to exist:

Bell Labs was made possible by a large-scale, vertically integrated telephone monopoly that allowed for an unusually long and wide research and development horizon for an industrial lab. Outside of those conditions (not likely to be repeated), funding a Bell Labs-style operation does not appear to be something most companies are willing to do. Even a company like Google, which spends billions on R&D and has displayed a willingness to fund speculative, longer-term moonshot projects like self-driving cars or life extension, doesn’t completely bite the Bell Labs bullet. Google’s Moonshot projects absorb billions in funding each year, but they tend to be organized as independent companies that raise money outside Google and get spun off when they seem promising enough. 

Bell Labs also took advantage of historical circumstances: discoveries in quantum mechanics yielded promising new phenomena, and WWII energized the organization while simultaneously creating scientific and technological progress that could later be capitalized on. These contingencies were the result of pure chance, not anything that could be controlled.

Part of corporations increasing unwillingness to fund unrestricted, speculative research likely comes down to the fact that labs with unrestricted research operations don’t appear to have been particularly successful. In Science in the Twentieth Century, W. Bernard Carlson argues that corporations turned away from unrestricted research because it didn’t achieve anything as transformative as the transistor or nylon:

In investing in R&D, American companies employed thousands of Ph.D scientists and built elaborate research “campuses.” At these new facilities, scientists were granted a large degree of autonomy, in the belief that such freedom had been the crucial ingredient in the development of nylon and the transistor. And yet despite ample funds, new facilities, and unprecedented freedom, scientists at the major corporate labs came up with few major breakthroughs from the 1950s to the 1980s. 

Looking at the history of the Bell Labs-inspired organizations seems to at least somewhat confirm this thesis. I haven’t studied their output exhaustively, but their most successful achievements — things like the Texas Instruments’ integrated circuit, IBM’s DRAM, and Xerox’s laser printer — mostly seem fairly closely tied to immediate product needs, and aren’t obviously the sort of thing you’d need to fund “basic” research to get. Indeed, research on the laser printer started years before Xerox formed PARC, and the other co-inventor of the integrated circuit was Fairchild Semiconductor, which as far as I can tell didn’t operate anything like a basic research lab. Their longer-timeline, “breakthrough” achievements seem to perhaps yield interesting scientific discoveries (the scanning tunneling microscope, Mandelbrot’s work on fractals, the SQUID), but not industry-transforming products. The only real “transformative” discovery to me seems like Exxon’s lithium-ion battery, and it was only subsequent developments outside of Exxon that made this successful.¹

There’s a well-known phenomenon that technological progress is often driven by bubbles, as irrational enthusiasm drives huge amounts of investment in a novel technology far beyond what can be economically justified. Carlota Perez describes the dynamic in Technological Revolutions and Financial Capital:

Two or three decades of turbulent adaptation and assimilation elapse from the moment when the set of new technologies, products, industries and infrastructure make their first impact to the beginning of a “golden age’” or “era of good feeling” based on them… Historically, those decades have brought the greatest excitement in financial markets, where brilliant successes and innovations share the stage with great manias and outrageous swindles. They have also ended with the most virulent crashes, recessions, and depressions…

It’s possible something similar happened with Bell Labs following the invention of the transistor. Not necessarily a bubble in electronics investment (though that could well be the case), but in the meta-idea of economically unjustifiable investments in pursuing basic research decoupled from immediate product needs. People often bemoan that American companies aren’t willing to fund research the way that they once did, but perhaps that age of industrial research was simply a bubble that was invariably going to pop.

1. Though it’s possible Exxon would have made these developments itself, had it continued to fund battery research.