Dr. Arati Prabhakar
Trade and Technology Seminar Series at the University of Maryland*
March 7, 1995
I am very pleased to be here today and particularly to be here with such a distinguished group. I wanted to talk about the role of technology in the global marketplace and to focus on the appropriate government role in that scenario. That topic is particularly appropriate right now. For example, the idea has been raised by some members of Congress that if we simply give corporations enough tax credits, they will provide for all the basic research in universities and build buildings on university campuses. I'm not sure which planet in the solar system that's going to happen on. It's my view that it is not going to happen on this planet. In fact I think the competitive environment for our industries is moving rapidly in quite the other direction and I would like to explore that today.
Let me start by talking about this business of technology. I found when I took this job in the Commerce Department, that instead of spending most of my time talking to technologists, all of a sudden I was mainly talking to people who were not technologists. So, I dutifully looked up the definition of technology in the dictionary so I could get it right. I found a fabulous definition that said that technology is the sum of the ways by which the members of a social group provide themselves with the material objects of their civilization. I really like this definition because it is broad and comprehensive. It says that it's the knowledge and the tools and the people, it's what's in peoples' heads and what's in their hands, it's all of those things that constitute technology -- and that technology is what we use quite literally to build our civilization. That's a very important way to start thinking of technology.
Beyond that, my personal fascination with technology lies in the fact that, to me, technology is a magic tool because it can change the ground rules. Technology is the single thing that can challenge fundamental assumptions about how we behave in our society, in business, and in the marketplace. Think for a minute about the many factors that drive a company's ability to compete in the marketplace: access to capital; the education and training of the available work force; the legal and regulatory environment that the company is operating in; and its ability to market, to distribute, and to trade around the world. All of those things matter tremendously, but technology is something that a company can grab hold of, can use to start changing the parameters of competition. Technology is the most immediate tool that a company or an industry can use to start changing the ground rules. Let me give you an example of how those ground rules can change in phenomenal ways that come from my own background of semiconductor technology.
There was a time at the end of the 1970s when people could see that the doubling of density of the number of transistors that you could put on an individual integrated circuit would continue for a long time. Gordon Moore projected that we would continue to see that doubling occur very, very rapidly, and every few years we would get that doubling. In 1979, people could imagine a time that we would be able to manufacture a million transistors on a single integrated circuit. And now that time has come, and we've even gone beyond that; we are to 10 million and more transistors on a single integrated circuit.
In the late '70s, if you looked at how we were designing integrated circuits and you extrapolated the number of design hours it would take to design something of the complexity of a million transistors, it started to look like it might take every man, woman, and child in the country to design every integrated circuit! It was pretty clear that wasn't going to work. Gordon Moore gave this fabulous talk in 1979 in which he concluded that the era of VLSI was coming, but that when we got to very large scale integration the only thing we would be able to make would be memories where you design each cell individually -- you don't have to manage a million transistors all at once. His conclusion was: "We will look back at the era of VLSI and all we'll be able to say is `thanks for the memories'" -- meaning, that we wouldn't be able to make any logic circuits.
Now, what I love about this story is that Gordon Moore today is the Chairman of the Board of Intel Corporation, which is the world's largest semiconductor manufacturer. They make a heck of a lot of money and they make it building 10 million transistor logic circuits, microprocessors. In fact, the reason that is possible is that when scientists and technologists looked at Moore's extrapolations of the design complexity and the number of hours it would take to design a complex integrated circuit, they realized that was an important problem to tackle. What came out of that, led by people like Carver Mead and Lynn Conway, was a design revolution that so dramatically changed design productivity, that today we routinely design million transistor circuits -- and we sort of sneer at those because what's really interesting is the design for 50 million or more transistor circuits.
That's the revolution that has happened in the business of semiconductors. Today, that same industry is grappling with capital costs for manufacturing. At all the meetings I go to where semiconductor manufacturing people get together and talk, the curve that everyone draws today is the one that addresses the cost of capital to build a new semiconductor wafer fabrication plant. For years we've been saying: "Oh, no. It's half a billion dollars, and it's going to be two billion dollars pretty soon." Today, it is two billion dollars. We're really right on that curve and we're still drawing these curves where the cost explodes. My contention is that, quite similarly, there are technology solutions that can change that reality, and that revolution is beginning to happen as well.
I know that in your own fields you can think of many similar examples, where technology took something that people treated as if it were written on stone tablets and really obliterated the assumptions, and really dramatically changed the future.
Well, what's new in the marketplace today and how does that relate to what's happening in technology? As I talk to hundreds and hundreds of companies in many, many different areas of industrial endeavor, there are several common themes.
First and foremost is the idea that today we live in a global competitive environment -- that markets are global and competition is global. That seems obvious today, but when you contrast it to the conditions under which our industries operated in this country in the period after the Second World War, I think it's quite dramatically different today. We tend to say: "Remember the golden days when all the American companies were on top of the heap?" That was a time when much of the rest of the world was picking up the pieces from war-time and figuring out how to move forward. It was a very fortunate era for us as a country, but today's reality, with fierce competitors in every part of the globe, is going to be tomorrow's reality as well. That competition will only continue to increase. It will not turn around and head in the other direction. There tends to be a lot of "wringing of hands" about how this makes life difficult, but it is that very competition that is driving efficiency in industry, that is driving costs down for performance. It's giving consumers tremendous access to new services and products. That is actually a very powerful force, however difficult it makes life.
We're living in a world in which product cycle times are getting shorter and shorter. There are industries, particularly in the information technology arena, where those cycle times today are measured in months, if not weeks. Among other things, that means that the number of years over which a company can capture the benefit of the innovations that led to that product is shrinking over and over again, and this is an important part of the increasingly competitive pressures for companies. Technology drives many of these trends. I specifically mention information technology's role, because my view is that today we are at the knee in the curve, if you like, in the information revolution. Information technology today is not just for computer scientists anymore; it's something that's beginning to affect every kind of enterprise in our society, and most particularly the business enterprise. It is very hard to find a company that is not touched by that technology. I think we have tended to think of high-tech companies as being separate from the vast majority of our industries. Today, every company is grappling with those changes in information technology. Again, it is pressure, but it is also opportunity.
A lot of these forces are driven by technology. At the same time, these forces that are driving the competitive approaches in industry are in turn are driving the way technology is developed. I think those trends have been very clear now for several years. If we look in an aggregate sense at what's happening in many, many industries across the board, I think there is a very clear trend that's been much discussed about the move to shorter term, narrower investments in R&D. Companies every year are working harder and harder to make sure that their investments in R&D are in those areas where that company can get the benefit, can appropriate the return, as economists say. Particularly in the science and technology community, once again there is a lot of wringing of hands and bemoaning this trend. But the fact of the matter is that this is absolutely superb business behavior for an individual company. It is a superb way to maximize profits, particularly short-term profits. It is a superb way to drive efficiency. So, I don't think that this is something that has to do with business people making bad decisions all of a sudden. In fact, I think the good news is that R&D is becoming a more efficient activity in companies, both in doing and in using it.
There is bad news that also goes with this phenomenon. When you start looking at what the benefits of R&D are -- because it's a knowledge intensive business, because any advantage that a company captures at some point will be available to competitors in the same business -- what becomes very clear is that the benefits to society as a whole or to the economy as a whole from R&D investments typically are dramatically greater than just the benefit that a company itself captures. So, if a company continues to optimize, to make just those investments where it alone captures the benefit, we really start to lose important opportunities in the economy as a whole over time.
Perhaps another way of thinking of this phenomenon is as a widening gap between the public-sector investments in basic science on the one hand, contrasting with the shorter term, product-development-driven investments that companies make. We need to think about that not in a context of: "This is bad for science and technology," but just that it is a fact of life in the global competitive environment in which our companies are operating.
With those comments to characterize what is happening with technology in the global marketplace, let me turn specifically to the role of the federal government. We need to recognize first and foremost that the power of our market economy is really quite phenomenal. When we talk about this business of technology and its relationship to economic growth, we are talking about a business that is owned by industry. It is industry's job to turn technology into products and services that lead to profits and jobs; there cannot be any confusion about that. It is also very clear that there are many things that the government has done for many, many years that affect industry's ability to do that job well. Obviously, many policies, particularly regulatory and tax policies, affect business' opportunities and ability to compete in a global marketplace, to innovate, and to do all the other things it takes to make technologies win in the market. That includes encouraging private investment, promoting work force, education and training, and opening up markets. All of those factors need to be addressed, recognizing the kinds of pressures and opportunities that our industries see in the global marketplace. None of these factors can be viewed in a vacuum. That's true for R&D and it's true for tax and regulatory policies. Even if we were to achieve the absolute optimal circumstance in terms of taxes and regulations, global competition and the pace of technological will not go away; they will only intensify. Given that reality, it is clear to me that corporations will continue to narrow their R&D focus. It's simply good business practice for individual companies. So none of the tax or regulatory solutions being talked about today is going to change the fundamental scenario of what's driving technology investment in this country and what our needs are going to be in the longer term.
I want to focus on one specific part of the overall federal role in our technology enterprise: the direct federal investment in science and technology, recognizing that it is but one of the many elements of the picture. We are all quite familiar with the substantial federal role in R&D in this country. About 45 percent of the nation's resources for research and development is provided by the federal government. It's a tradition that goes back at least 50 years, since the end of the Second World War -- although if you look back to the very beginnings of this nation, you will find some early versions of those same kinds of investments. Some of the discussions we are hearing today make it sound as if this is a new and radical idea. It is actually a very old idea that the government has a role in R&D.
I want to examine what that role has been and then to talk about what it needs to be in the current global competitive environment. Since the end of the Second World War, there have been two major emphases in the way we've invested federal R&D dollars. One was focused on basic science and engineering, really providing for the university infrastructure in this country. That was a role that was recognized and defined in the post-Second World War era and it has been one of the most powerful investments we have made in this nation. The fact that the United States has some of the richest science and technology capabilities in the world is largely due to that sustained, thoughtful federal investment and that coupling between the federal government and the university research system. That is something that is precious and very much under scrutiny and pressure -- but it's something where we have a very strong track record and have provided tremendous value to society.
The second major focus in federal investment has been to support science and technology for specific government-owned missions, primarily national security, but also space, energy, and health, as the major areas. As a consequence, R&D spending is distributed across many, many different government agencies, with the major dollars coming from the Departments of Defense and Energy, National Institutes of Health, and NASA.
Today we're talking about how we steer that federal R&D investment for maximum leverage for our country in today's environment. It is worth taking a moment and recognizing again how valuable that investment has been in the past. It is not something that we want to abandon; it is not something that we want to throw out the window. In fact, I would argue that it has been one of the most powerful of all federal investments in terms of the return to society.
But the world today is different, and I don't think that the models that were developed in a period when our situation in the world was quite different are adequate to do the job that is at hand today. As I look at today's technology policy more specifically, I see something old and something new. What is old, and needs to continue, is a recognition that the appropriate role for federal investment is not in the marketplace, not doing product development, not delivering technology to the market; that is what industry does extraordinarily well in this country. Rather, the federal role has long been to focus on the infrastructural technologies that all companies need but no one company benefits from investing in them. The federal government has focused on investment in long-term and risky science and technology because of the under investment of private funding. That's what hasn't changed.
What has begun to change and should change more purposefully are the mechanisms by which we tackle those jobs. Let me characterize our old model as saying: "We will fund basic science and research in the universities and we will fund mission-related projects, space for example, and it will be nice if these things become useful to the economy."
We could argue about whether that was the most effective model in the '60s. I definitely don't believe it is a sensible model in the '90s given the pressures companies are under in making their own investments and given their limited ability to bridge to markets from the much broader base of public-sector support for basic science and infrastructural technology. So, what is emerging today as we start to make some of these shifts in federal R&D is really a model of technology partnership programs. In this category I put not just the work that we're doing in the Department of Commerce like the Advanced Technology Program at my agency but also efforts like the emphasis on dual-use technologies in the Defense Department and the Technology Reinvestment Project. I would put in that category efforts like the Cooperative R&D Agreements, or CRADAs, that many of the federal laboratories around the country are using, and the Small Business Innovation Research awards, as well. The National Science Foundation's Engineering Research Centers can be thought of as part of this concept, too. I think it is essential that we build those bridges, because otherwise we will continue to spend our $70 billion a year on federal R&D but it will not be as connected and as relevant to our economy as it needs to be to serve our society. We also must recognize that this is not about moving that entire $70 billion investment into the business of technology partnerships, because we simply cannot abandon the science base. That is an important responsibility. We can't abandon mission-related R&D. That is very important as well.
What we are talking about is making a small investment to build that bridge from basic science to the technologies that will power our industries in the years ahead. Today, if you sum the funding for all of the programs I mentioned in the category of partnership programs, you will come up with less than five percent of the total federal R&D investment. NIST's investment, our total budget today, is just over one percent of all of federal R&D spending. So, again, with that as context, I think it is important to make sure that we invest in that bridge with some fraction of the resources.
At NIST, we are working very hard at this job. But we are also being very careful to make sure it is sensible and that it does work. I've been talking about these ideas as if they were brand new and sprung full blown on the horizon in 1988. But the fact is that there has been an enormous amount of thinking and planning and trying to get these kinds of activities to happen over literally decades as people recognized the changes that were going on in global technology competitiveness. There have been some false starts and a lot of lessons learned. We have tried to build on that experience and use it to structure our current experiment so that it will succeed. And while it is still very early, my belief is that we are beginning to succeed.
I wanted to talk about the elements of this experiment. There are four key characteristics that are the underpinning principles for our investments at NIST and this business of technology partnerships.
First, we recognize that the appropriate role for our funding is for infrastructural and long-term or risky R&D -- and we recognize that the federal government's role should not be to take technology directly to the marketplace. It should not be to do product development, marketing, or production. We recognize the federal government should not be providing land or building factories, or subsidizing dying industries -- all of which other industrial nations actually do. We just don't think these make sense in the U.S. economic structure.
Second, we recognize that our mission, focused on economic growth, is one that is not owned by the government and, therefore, is not sensibly addressed by all of us sitting up in Gaithersburg and deciding on the right answers. We know that industry needs to drive the priorities, that it needs to be industry's view of market pressures and opportunities that points the directions. That is a critical element in everything that we do.
A third principle of our programs focuses on competitive processes: recognizing that when the country has tried to do these kinds of programs in the past, on a political basis, it has not been a successful investment for taxpayers, however successful it may have been for the recipients of the funding. We are trying to model our approach on a long history of peer review in the science funding domain, recognizing that competition on a merit basis leads to the best investments for taxpayers.
The fourth principle of our approach, thinking of this as a grand experiment, is focusing on evaluating these operations so that we can make ourselves more and more efficient. We focus on some of the near-term outcomes to keep the efforts on track. But more fundamentally, we are focusing on assessment of the economic benefits that flow from the technology investments that we are making. It is essential for us to figure out if the experiment is working, it is essential for us to do the real-time process control that improves the experiment as we go along.
I would be happy, as I'm sure several of you know, to tell you for hours at length about our individual programs and some of the others across the government. But I really want to stop at this point, having put some of these ideas on the table and turn to the discussion. Building these bridging programs, these technology partnerships, and learning to cooperate does take a long time, and it is sometimes painful. But in a handful of years we've made dramatic progress. It is too early in this experiment to see the complete, full-blown economic impact that flows from these kinds of investments. But we do have in hand today the kinds of very powerful early indicators that tell us that we are getting the kind of business activity today that leads to economic impact tomorrow. We need to acknowledge that if we are going to make long-term investments, there also will be a long-term evaluation period before we can really tell what the results are.
In the political discussion that is going on today what is at stake is the idea of government and industry partnership, of building a partnership that connects to the wider technology community of universities and labs around the country. The idea that there are technology jobs that matter for the nation. That we can't do everything separately and we must do some things together. That is the idea that is at stake and on the table today. We are going to be very active in continuing to build that discussion, hopefully turning it into much more of a debate. We are going to continue to advocate for moving in this direction of partnerships, and I really welcome the opportunity to hear your questions and comments as we try to elevate the quality of conversations about this very vital subject.