Edison’s Legacy: Industrial Laboratories and Innovation by Henry Kressel see note please
David “Spengler” Goldman writes: Dr. Henry Kressel’s essay on industrial laboratories may be the most important piece of economic writing I have read in the past several years. It’s a practical guide to restoring American technological prowess and productivity growth.
Between 1999 and 2016, the U.S. share of global high technology exports dropped from 18 percent to 7 percent. From one of the world’s leading technology product exporters prior to 2000, the United States has become a net importer since then, and the deficit keeps growing. During this period, China’s share of exports increased from 3 percent to 26 percent, reflecting the shift of manufacturing overseas—including important high technology industries that were pioneered in the United States as early as the 1950s—suchas telecommunications systems, consumer electronic products, microelectronics, and solar energy converters.1
While we have seen the emergence of innovative U.S. companies in information industries, of which Google is a prime example, we have not seen a similar development in technology sectors involving domestic manufacturing. In fact, the sale of Westinghouse’s nuclear energy division to Toshiba is an example of a domestic loss of control over an industry vital to the United States Navy (though the company now appears likely to be sold back to U.S.-led private equity consortia). Or consider another example among many: key components of advanced computing technology essential in military systems, such as custom processing chips, may be designed here but are manufactured in Taiwan.
The domestic decline of important and vibrant technology industries impacts economic growth and contributes to the loss of well-paid manufacturing jobs. It also adversely affects military and other technological capabilities. The loss of a domestic manufacturing base for vital industries means that continuing innovations in those sectors are difficult to create or control. Computer displays, for example, are produced overseas, and that is where the innovations are now being developed and commercialized—because core innovations need to be embedded in the manufacturing process. While the migration of industries from developed to emerging countries is a common historical tendency, the shift away from the United States in recent decades has been unprecedented in its speed. To solve this problem, the United States must increase the rate of domestic industrial innovation and secure the domestic base of advanced industries. This calls for a major initiative involving industry, universities, and government over a period of many years. Every year we fall further behind makes it harder to recover.2
There are precedents for successful U.S. national efforts to boost critical technological innovation, such as the space program under President Kennedy and the ballistic missile defense initiatives (“Star Wars”) under President Reagan. With combined federal and corporate commitments and funding, remarkable progress can be made in moving breakthrough innovations from concept to product—and in the process creating major new industries.
Replicating those big innovation programs today, however, calls for a different execution strategy, because the industrial landscape has changed. The biggest change is the disappearance of corporate laboratories, which were part of large companies and had funding that allowed for long-term projects with potentially big impacts. These labs also brought together interdisciplinary teams of scientists and engineers for extended periods. Furthermore, the parent companies of these laboratories had the resources to move concepts into the market. The point is that maintaining a leading-edge economy the size of the United States requires combining the skills of the most talented people with appropriate resources to build market leadership, and corporate labs were critical components of this process.
America once had a few large, well-funded, and well-managed multidisciplinary corporate laboratories that housed some of the most brilliant technological researchers. They worked in environments where exceptionally creative people could innovate and see the fruits of their work translate into breakthrough products. A major virtue of such labs was that unexpected product ideas could emerge as researchers followed their curiosity to discover new phenomena. New materials and devices were invented without the pressure to produce quick results or to work only on low-risk, evolutionary product development—the typical task of most engineering departments associated with product divisions in corporations.
Such large, interdisciplinary corporate labs were the primary generators of new electronic and materials technologies. In fact, most of the basic innovations in computers, semiconductors, and software that enabled digital technology came out of the big U.S. corporate R&D organizations, which were formerly part of AT&T, General Electric, IBM, Xerox, RCA, and a few others. In addition, start-up companies that grew into giants such as Intel or Cisco initially leveraged technologies developed by the big companies or university laboratories. For example, Intel leveraged semiconductor technology from Bell Labs and Cisco originally leveraged digital communications technology developed with defense funding.3
But these big laboratories no longer exist. There is a great deal of innovative work in the United States—in “big data” and related information technology such as “machine learning”—but where are the institutions that will foster the next generation of breakthroughs in process technologies, new materials, or infrastructure technologies? This is the subject that I address here—with opinions shaped by my experience for over twenty years as a researcher and a leader of one of the great central corporate laboratories, that of the RCA Corporation.
Edison’s Legacy: The Industrial Laboratory
How do breakthrough industrial innovations begin? The twentieth century has often been called the Age of Edison, in recognition of his role in creating many of the technologies that have shaped the modern industrial world. With 1,093 patents to his name, Edison has been held up as the archetype of the lone inventor, producing new innovations by sheer force of personal genius.
This is a myth, however. Thomas Edison did not work alone. He may have been the driving force behind his many inventions, but he was supported by a handpicked technical team assembled for the purpose of inventing new technologies. He built the first industrial laboratory, designed to turn radical new ideas into marketable products.
The composition of his team marked a change in attitude toward scientific endeavor. Since the beginning of the Industrial Revolution there had been a sharp cultural distinction between “pure” scientists, who advanced the state of knowledge, and “practical” inventors of commercial products. Scientists enjoyed high status, while inventors were considered a lower order of commercially minded technicians because they did product development.
The recognition that this cultural gap must be bridged within a single industrial organization is one of Edison’s key contributions. In 1876 he created the world’s first industrial laboratory in Menlo Park, New Jersey. This organization, which he dubbed the Invention Factory, successfully integrated scientists with technologists. His lab made history: the phonograph, motion pictures, incandescent lighting, and electric power generation (among many other breakthroughs). What is equally impressive is the number of inventions that eventually created entire new industries.4
Edison’s success encouraged others and led to the formation of several large industrial laboratories to provide an environment for far-reaching innovations. Their funding remained mostly steady to protect long-term programs from the vagaries of business cycles. The funding also assumed that the product divisions would have the ability to exploit the innovations developed in these laboratories for commercial gain.
The most important electronics research laboratories were part of vertically integrated corporations covering broad industrial sectors, such as American Telephone and Telegraph (AT&T), International Business Machines (IBM), the Radio Corporation of America (RCA), Westinghouse, and General Electric. AT&T’s Bell Laboratories was founded in 1924 and engaged in work ranging from basic research through product development. By the early 1980s, when AT&T was broken up by agreement between AT&T and the U.S. Department of Justice, Bell Labs employed about twenty thousand scientists, engineers, and support staff distributed in various facilities around the country. Long-term research (by a few hundred researchers) and product development by a much larger number of people could be funded over multiyear periods, because Bell Labs benefited from a unique industrial advantage—it was part of a government-created telephony monopoly. This happy set of conditions made it possible to support research and development covering all aspects of science and technology related to communications, and to provide a path from research results to products, all within AT&T. Innovations accomplished over decades at Bell Labs made possible the modern wireline and wireless telephony systems and included noteworthy contributions to computer science as well.
Many other corporations maintained R&D labs, but a few stand out. These labs attracted the most talented PhDs in science and technology. By offering outstanding researchers first-rate facilities and the ability to pursue programs that took years to complete, these laboratories contributed to many of the great innovations that have shaped the modern world. Here are some of their most noteworthy contributions to electronics, many of which eventually created new industries:
- In addition to major contributions to computer technology, IBM’s research laboratories were early pioneers in electro-optic devices as well as digital data mass storage technologies.
- Westinghouse Research Laboratories contributed active matrix liquid crystal displays—the foundation of all flat-panel displays.
- Hughes Research Laboratories developed the first practical gas laser.
- The Xerox Palo Alto Research Center developed the visual computer interface, the Ethernet protocol, and the laser printer.
- Another major contributor, RCA Laboratories, will be discussed below. I spent over twenty years there, first as a researcher and inventor and later as head of corporate R&D of electronic devices and systems.5
Evolutionary technology development is done in many engineering departments, but the breakthrough innovations that changed the world are not evolutionary ones—their birth requires very special environments where creative people have the freedom to follow their instincts and where management is focused on longer-term value creation.
Let me dwell on the RCA experience for illustration. RCA Laboratories, founded in 1941, first made its reputation by inventing every one of the core technologies (including manufacturing methods) for color television by 1950. NBC, owned by RCA, launched color TV service in 1951. The development and successful commercialization of color television by RCA is an excellent example of the value of innovative research in a vertically integrated company, driven by a visionary leader, David Sarnoff. No independent research organization could have succeeded without RCA’s access to capital, its broad technical and scientific skills, and its control of a television broadcast network.
An important element in the success of RCA Laboratories (and other successful big, central labs) was a management style and culture that merged individual initiative and corporate needs. My experience convinced me that creative people will focus on solving important industrial problems when they can do so in an environment that recognizes their achievements and the broader impact of these technologies. In fact, talented people may prefer an industrial environment over an academic one because of the satisfaction of seeing their work impacting the world.
That is why RCA Labs management encouraged researchers to learn about industrial problems and participate in moving their innovations into products. The response varied—some individuals were best suited for academic research—but some researchers took up residence in RCA product divisions during the course of new product introductions based on their ideas. When the product launch was completed, they went back to the labs and started other projects, applying the valuable experience that they had gained, plus an appreciation of what it takes to deliver a product. They derived great satisfaction from seeing their inventions make a direct impact.
The endless task of lab management was to promote creativity, keeping business objectives always in mind. To promote creativity among researchers, young PhDs were encouraged to define their own areas of interest. My experience offers an example. After completing my PhD, I decided to research semiconductor lasers, then considered useless laboratory curiosities because of their poor operating characteristics. Given the freedom to experiment, I quickly discovered why the lasers operated poorly, and I co-invented a new heterojunction architecture for the devices and developed a better manufacturing technology, which was transferred to the RCA Solid State Division. These new lasers lasted for many thousands of hours and eventually found their way into the most demanding communications systems, and even into air-to-air missile guidance systems. In just two years, this revolutionary device went from laboratory curiosity to the first broadly useful commercial semiconductor laser, introduced by RCA in 1969. This rapid effort was only possible because the lab’s organization and staff were so flexible; talent needed for specific programs could be assembled very quickly.
Similar successes were achieved many times throughout the history of RCA Labs. During the 1960s and ’70s, RCA Labs (renamed the David Sarnoff Research Center in 1951) pioneered such important devices as liquid crystal flat-panel displays, high power silicon transistors, thyristors, and rectifiers, the first metal oxide semiconductor (MOS) transistors, the first CMOS devices, the first practical solid state imagers, solar cells, semiconductor lasers, rewritable optical storage discs, the first magnetic core memory, and a host of new microwave devices.
Not all of the work was R&D—divisional support was ongoing, and the staff consisted of more engineers than scientists. When other organizations within RCA did not have the talent they needed, the labs stepped in to complete prototype systems and products. For example, we developed key components of the radio that connected the Lunar Orbiter with the Lunar Explorer Module on the moon during the 1969 Apollo mission, which used a microwave power device that I developed. Another example: the labs developed the first lightweight portable TV-quality color camera used by mobile TV crews, built with the first solid state imager. It was initially used by NBC—another example of the value of being part of a vertically integrated company. The achievement won RCA Labs an Emmy award in 1984.
Why Central Research Laboratories Disappeared
By the late 1970s the central labs were in decline primarily because foreign competition put pressure on profitability and because corporate structures and strategies were rapidly changing. The enormous effort to rebuild the industrial base of Japan, starting in the 1950s, soon produced a formidable competitor.6 Japanese companies, with the collaboration of their government, had organized themselves to achieve leadership in selected industries—and electronics were at the top of the list. By the 1970s, the full force of this competition began to hit American corporations.
Impacted companies mobilized all of their resources, abandoned loss making products, and cut costs by moving manufacturing out of the United States. Research funding for long-term projects was often the first victim: product division management had to make short-term decisions on their strategy to counter competition. Under this sense of urgency, central research laboratories and their longer-term projects looked like a waste of money, so staff was reassigned to short-term, product-specific projects expected to produce short-term paybacks. As a result, some of the most creative scientists with established reputations left to take academic positions, and as a result universities such as Stanford acquired highly experienced research professors.
As corporate funding waned, the Department of Defense stepped in to support these laboratories and leverage their talents with increased funding for electronics research. This led to the development of new technologies for military applications, producing many different kinds of semiconductor devices now used in computers, communications, or microwave systems (radar). The Defense Advanced Research Projects Agency (DARPA) and the National Space and Aeronautics Agency (NASA) were the major funding sources for electronic devices and systems of all types, which eventually found their way into commercial products—in many cases by start-ups. For example, the CMOS chip technology universally used today came about because of the DoD’s need for power-efficient chips not then available through commercial technologies. The laser development that I headed was similarly a recipient of DoD funding to enable fiber optic communications systems.
But income from DoD contracts was ultimately not enough to sustain the central laboratories. Such organizations gradually became engineering contractors for product divisions and lost their ability to conduct long-term projects. Their funding became tied to specific product objectives set by the divisions.
By the late 1980s, the big corporate electronics labs with significant long-term research programs were largely history. The 1984 breakup of AT&T split up Bell Laboratories, and the pieces ceased to be viable contributors. When General Electric acquired RCA in 1984, RCA Labs (later renamed Sarnoff Corporation) became an independent subsidiary of SRI International, a major government-funded not-for-profit research institute. The Xerox Labs were also spun out as the company abandoned its diversification strategy (and the interest in broader electronic products) to focus on printers and copiers. Both the GE and IBM labs were downsized and largely abandoned broad-gauge research.
As the big labs disappeared, or changed their mission, electronics research became largely the work of universities. This is reflected in the contributions to scientific literature. Between 1988 and 2004, corporate lab paper contributions to Applied Physics Letters, a leading research journal, dropped from 40 percent to 5 percent, while contributions from universities increased from 45 percent to 70 percent.
Achieving Innovative R&D: Lessons from History
Can (or should) the old model of big, interdisciplinary corporate labs be resurrected? This is highly unlikely, even if we increased R&D funding. One of the strengths of the central laboratories was also a liability. Moving into new areas where technology was not available was difficult. The expertise they had built up over many years in selected areas of technology was not necessarily what the corporation needed to rapidly satisfy new, changing market needs. It requires time to hire and train people, and making corporate acquisitions to move into new products seemed to offer a quicker path to market.
More flexible organizations are preferred today, and this is the subject I now address. Any successful strategy will need to allow for demands of increased flexibility, without ignoring the lessons or advantages of the large corporate labs. Over their lifetimes, the leading central laboratories produced remarkable breakthroughs, so what are the important lessons from their operations that are still relevant—and may be missing—today?
(1) It starts at the top. Risk-averse and ignorant corporate management cannot be expected to foster or bring to market breakthrough innovations. Innovative organizations require senior leadership that understands risk and knows how to manage it in its chosen industrial environment. The management of creative people to meet industrial objectives requires an unusual set of management talent—of which David Sarnoff was an excellent example—which was a distinguishing feature of the leading corporate laboratories.
(2) A select number of the most creative people should be given the freedom to follow their instincts. For this effort to be productive, careful selection of team members is required. In my experience in managing such organizations, I found that the best contributors were scientists and engineers who have demonstrated unusual levels of initiative in an engineering organization. They were people driven by the desire to create and make their reputation by the success of their creations—not people just attracted to research in an abstract sense. Such people define new product ideas and have the mental and organizational skills to achieve practical results.
(3) Selecting and scaling major projects from the many ideas that are generated is critical. The decision to proceed with product development should be made by an informed group that understands the expected costs and opportunities of novel products or services. This selection committee should be small (three to five people) with experience in the sector and prior product development history. That selection process is critical because it is always easier to say no than to incur new development costs on risky programs. I am not talking about a single committee for all projects, but a number of committees devoted to various sectors that require deep knowledge and experience. It sounds difficult—but growing innovative companies always is.
(4) Once a decision is made to proceed with a potentially breakthrough new product, the concentration of multidisciplinary scientific, engineering, and marketing skills in one corporation speeds the path to market.
(5) How funding is managed can determine the success of programs. Short-term funding is appropriate for early feasibility work, but programs need a mixture of funding sources that allows long-term programs to be conducted separately from short-term product milestone funding.
To implement the above strategy in the absence of multidisciplinary laboratories, corporations are experimenting with more flexible organizations. Often called innovation centers, the goal of these organizations is to move from concept to product by promoting creative thinking at the working levels.
Innovation centers are dynamic organizations within corporations that focus on specific technology areas.7 The number of people involved may not be large, though the size of the teams should change with the scope and skill requirements of the projects. But the idea is to give creative individuals the freedom to develop new ideas, initially providing limited budgets to prove concepts and then moving to product implementation on the basis of promising early results.
For such innovation centers to be effective value creators, there are a number of important guidelines. Such a center should be entirely dedicated to creating new product concepts. It should report to the CEO or corporate head of R&D and have a separate budget. The corporation should incentivize business units to move innovations to the market and to defray costs, while also recognizing specific employees who contribute to commercially valuable innovation.
The projects within an innovation center must have breakthrough product potential. The value propositions of these projects should be developed from concept through commercialization. In some cases, completely new business units might eventually be established around a core innovation center team if the anticipated products do not fit within existing product organizations. Once a promising product concept is generated, its commercial implementation requires that teams be established from all corporate units involved in the commercialization process. The team members should come from the innovation center, be matrixed or assigned from the operational divisions, or be recruited from the outside. The teams require experienced leadership that has the respect of the organizations involved in the transfer to the market.
Most managements hate change. Conventional product line management is often faced with the difficult task of prioritizing among projects: the development of near-term products demanded by customers; evolutionary product development initiated by their own engineers; and riskier long-term projects, which they view with more skepticism than those with shorter horizons. RCA’s approach to break down this barrier was a budget for innovation projects, which was provided to business units to help defray the cost of getting a new product through final development and into production.
An Example from Israel
The above concept of flexible innovation teams has been implemented. The creative ideas of a small team led to the creation of a revolutionary defense product in Israel at the Rafael Advanced Military Systems Company. This profitable company with multibillion-dollar revenue is owned by the Israeli government but operates as an independent, for-profit business exporting innovative military products. Israeli start-ups receive considerable attention and admiration, but most of Israel’s important military technology comes from big, multidisciplinary manufacturing companies such as Rafael. And the example of the Iron Dome short-range missile defense system shows how it happens.
A small group of engineers who had worked on aircraft technology came up with the idea of applying the principles learned to short-distance, low-flying missile ground defense. They put together an exploratory feasibility study program funded by management and the Israeli defense ministry. When the results proved promising, the ministry funded the next stage of development, bringing into the program the interdisciplinary skills needed to build a prototype and conduct a field test. As the success of the program continued, more resources were added in order to get to the next step, but the project remained autonomous until the production stage. A key factor in the success of the program was the personal support of the Israeli defense minister, who was its champion against many detractors who did not think that the system could work.
In short, this structure allows for flexible but high-risk development. Within the engineering organization, Rafael encourages “ground up” ideas that can be the seeds of breakthrough products. And once a promising concept is demonstrated, resources are pulled in from various engineering and production departments to achieve a product result. Managing such a process obviously requires a great deal of discipline, but this model has the advantage of providing both flexibility and sufficient resources, while motivating staff to innovate and gain the satisfaction—and recognition—that come from participating in major new developments.
U.S. Technology Priorities Must Be Established
As noted earlier, many industries requiring manufacturing investment, even those based on American innovations, have largely moved overseas. The result is that leading-edge technologies essential for ensuring American defense superiority are at risk of not being available domestically. Advanced communications products and semiconductor memories, for example, are now produced largely in Asia. Addressing this problem requires that the Department of Defense conduct a thorough analysis of technology requirements and capabilities (extending some years into the future) and establish priorities. Prioritized programs then must be pursued with appropriate funding—extending from research and development to manufacturing and production. Various agencies such as the National Security Agency have technology development roadmaps, but it is necessary to fully map the future requirements that impact all parts of the defense effort.
We also need to examine our ongoing defense programs in the context of these needs. On that basis we can set priorities and select institutions for funding and managing the transition from concept to product. This inevitably means tight collaborative programs between corporations, universities, government labs, and independent research organizations. In this regard, DARPA has historically played a key role in managing the funding of leading-edge technology development (including microelectronics, lasers, and the internet, for example). This organization has demonstrated great skill in working with the various parts of the Department of Defense in establishing priorities and managing funding, and its role should be expanded to lead this larger strategic effort.
Because many technologies and sophisticated systems are best developed in multidisciplinary environments, organizations selected for funding must be of sufficient size and adequately equipped to turn concepts into reality. As discussed earlier, the model of innovation centers offers a basis for building creative corporate environments. To that end, it is important to fund centers of innovation for selected technologies that will attract top talent and be provided with sufficient resources to manage a transition to product deployment (in collaboration with appropriate partners). To ensure that the transfer to market happens, government funding may be needed in some instances to offset the costs of turning marketable products into military systems—particularly in process industries associated with manufacturing.
Universities are key participants, too, but their work needs to be linked to practical ends once the feasibility of new technologies is demonstrated. With the demise of basic and applied research in corporations, university researchers have performed seminal work with government funding. Yet while these activities have led to a great deal of creative work, there is a severe bottleneck in the transfer to useful products.
That transfer process is critical in building value from innovative concepts, and it requires industrial development and investment. There is no substitute for sophisticated product development work performed by engineers familiar with practical requirements. A successful example is SRI International (the former Stanford Research Institute), which led an R&D software program in machine learning development. This program received over $150 million of funding from DARPA and involved a consortium of universities, national laboratories, and corporations.
In addition to military applications, one of the results of this program was Siri, the popular iPhone feature (Apple acquired it from SRI), and copies have hit the market from other companies. But the target for selection of DoD funding priorities should be based on defense needs. History has shown that subsequent commercial outcomes follow once the risk of proving the value of breakthrough innovations has been validated with DoD funding.
Finally, a word about private venture capital funding. In general, such funds favor capital-light start-ups based on software businesses with relatively low infrastructure costs, not companies offering new technologies for manufacturing sophisticated electronics. For example, in 2016, venture capital investments in information technology software was $1.8 billion, while investments in information technology hardware companies was $230 million. Capital intensive process technologies inevitably have fewer venture capital funding opportunities. Established corporations with more resources are the better outlet for such developments.
The United States is in the perilous position of losing its technological edge in critical manufacturing industries—including those that underpin leading-edge defense technology. Historically, DoD funding helped create many of the technologies that underlie the modern world. These technologies also made the United States an international leader in advanced commercial and defense industries.
If the United States is to continue fostering the kinds of technologies that create both economic prosperity and defense superiority, then a concerted effort to support research and development is needed. The development and implementation of transformative technologies require the collaboration of corporations, universities, governments, and R&D laboratories, all focused on bringing innovations into practical use.
There was a time when big corporate laboratories were at the center of such innovation. But those labs are gone. And although they are not coming back, certain of their core functions must be replaced. New, more flexible organizational structures, staffed with the most creative scientists and engineers, must be created and funded to achieve the long-term technological breakthroughs needed to ensure U.S. economic and military leadership.
Comments are closed.
