appropriately connected within the organization to do that. We must have strong relationships with development and manufacturing, which we did not always have.
Second, we build and apply a technical base across a very broad spectrum of physics, chemistry, materials, mathematics, engineering, and computer science, sometimes building it without a clear, detailed view of how it will be applied. It is important to determine what is relevant and what is not, and to build the science underneath that.
Third, in our business, if you look back over a period of five years and examine where the growth in the industry has come from—we are in an industry that is worth about $600 billion now worldwide—you find that half the growth has been generated by technology or products that were not there before. So, if all you do is extend your mainstream technology, your business is probably going to shrink. Part of our job, therefore, is to provide the opportunity for new business. The normal reaction of people who are in a particular business, when you come up with a new idea, is that it sounds strange, that it is risky and that it might even be a threat to the business. You need some organization like the research organization to pursue new ideas. A good example is parallel computers that are not tied to mainframes, an innovation that resulted in IBM's SP-1 and SP-2. In that case, the research division played an important initiating role and made strong arguments, which were eventually accepted, for the need for such stand-alone parallel machines.
Then, finally, the world we are in is very science based. You have to have a set of people looking around, trying to let nature tell them what can be done next. What is the next display technology? What is the next way you can store bits? What is the next way you can advance some communications technology? You have to start at a very fundamental level, for instance with people studying materials, physicists who are in close contact with the peer research environment, who can recognize the importance of certain advances, and can then make the transformations necessary to focus those advances in appropriate directions.
So, this is the inside-out view. Another thing that we have done over the last five or six years—and it was driven largely by our focus on quality—is to start from the outside and work our way back in. What is this company about? What kind of businesses is it in? What is the customer's view of value? We do not leave it up to the company to figure out what that value is; we take the lead in figuring it out and then explain it to our sponsor. This is a very fundamental principle of research organizations. If you want to survive in the long run, you must understand what value is—what value is in the eyes of the sponsor. Researchers have to do that type of thinking and explain it to the sponsor.
It is also important to recognize that value can change radically. For example, the Cold War ends suddenly, and defense technology is no longer highest on the list of priorities—economic performance is higher. That directly affects what value is and must affect the resulting research agenda. For NSF, anticipating this kind of thinking about value is key. The law that we are discussing today is a good one, but in thinking about value, measuring how well we achieve it is necessary whether the law exists or not. And we are better off being out in front, not in a reactive mode, explaining value only after being asked.
I am quite sure that most of you are aware of the kind of problems we have experienced at IBM in recent years. After all, we lost $8 billion in 1993. Over a period of three years, we lost $18 billion. The chief executive officer (CEO) was replaced. A new CEO, Louis Gerstner, was hired in April 1993, and one of the first places he visited was the research division. He liked what he saw because we had begun this change before we had budget problems, before the company had financial problems, and before anybody in the company was pressuring us to make these changes.
Figure 2.1 dates from around 1991. IBM is in three kinds of businesses. We are in the integrated solution business. That means, for instance, that we work with Bank A to help it compete with Bank B across a whole range of technologies—to help it translate technology into business value. Second, we have products. These are the source of most of our revenue today—software and hardware products, mainframes, personal computers (PCs), communications products, disk drives, and that sort of thing. Underneath that, we have a set of technology businesses, which were not always businesses. In an earlier time, they were merely suppliers to the internal customers. So, our semiconductor business did not sell anything outside. This situation has changed over the past few years. Last year, we sold more than a billion dollars worth of our semiconductor products outside IBM. That, in itself, relates to quality; contact with the outside world helps ensure our competitiveness.
Now let us turn to the outside-in view of value—value as seen and measured by the customer, the sponsor. For us, this means the measure of value as seen by our chairman. We have identified seven aspects of value, six of which are listed in Figure 2.1.