Mr. GROVE. With that, I would like to depart from that written statement a little bit, rather than follow it point by point, and try to highlight the exact messages that I am trying to convey there, if I may do so.

Mr. KASTENMEIER. Actually, Mr. Grove, you have a very short statement. We will be pleased to receive it for the record, together with its attachments.

Mr. GROVE. By way of introduction, it is very generally recognized by those of us in the industry, in the industry in a broader sense in the electronic and high technology industry, that essentially everything that has happened in the last two decades in terms of electronics and high technology endeavors have been fueled by the development of the semiconductor business, semiconductor industry, our industry, the industry you've heard described.

The driving force for these developments has been the phenomena of absolutely, steadily, predictively dropping prices, and dropping costs of electronic functions. Mr. Sevin used electronic memories, the capability of storing numbers, as an example. The cost of storing one number that we refer to in the lingo of the industry as one bit of memory, has dropped by something like between one hundred and one thousandfold in the last 10 years. There is no magic involved with this. We operate in the same inflationary economy as everybody else. We use a certain amount of automation, but it's a very, very miniscule contribution to this cost reduction. The basic method for this gigantic cost reduction has been, and continues to be achieved, is putting more and more functions into a little package like you have on your desk; putting a larger and larger number of functions, bits of storage into the same package.

At the risk of some oversimplification, if you look at that package, a package like that has always cost about the same amount of money to make, roughly. Ten years ago or twenty years ago there was one transistor in it, and it cost less than $20 to make. Today, there are 30,000 to 40,000 transistors in it, and it costs about $20 to make.

Mr. KASTENMEIER. Quite apart from content, Mr. Grove, does the industry generally follow the industrial processes described by Mr. Sevin, in fact illustrated by Mr. Sevin? Because the steps look very complex. It would not necessarily follow that everybody used precisely the same technology to achieve the same results. Is there much difference in industrial processes?

Mr. GROVE. Between one company and another?

Mr. KASTENMEIER. Between one company and other.

Mr. GROVE. There is always some difference, which is why some companies do better in certain product areas than others. The differences relative to the complexity you saw displayed in the slides are minor, but at a given time those minor differences are somewhat corresponding to a racing car turning a corner or a sedan turning a corner. They both turn, they both have four wheels, but one of them is going at the edge of its ability. The differences are kind of subtle. If you look at them far away, they are both automobiles. In that sense the processes are always different. But all competitors seem to be marching along and improving their capabilities as time goes on.

Mr. KASTENMEIER. Are those processes in part or in whole covered by patient protection?

Mr. GROVE. In part they are covered by patent protection; in whole, they cannot be. There is a great deal left that is unpatentable and sometimes it is better served by not being patented.

If I may continue:

The flip side of this gigantic increasing technological complexity as we put a larger and larger number of functions into a given package, is that the cost of design has literally skyrocketed. It has not increased in proportion to the number of components that you are trying to fit in. Just like a jigsaw puzzle, the time to take to put a jigsaw puzzle together increases more rapidly than the number of pieces in a jigsaw puzzle. You try to put a 100-piece puzzle together and a 500-piece puzzle together, the 500-piece puzzle will take more than five times to put together. It is a very analagous process. The design process that Mr. Sevin described is very much like putting jigsaw puzzles together, because you have to put these components in close proximity to each other, have to make them fit.

Because of that, increase in the time taken to do design, integrated circuit design, has absolutely skyrocketed. I have a real-life illustration, it so happens, from the keynote speech at the most prestigious industry-wide conference that took place 2 months ago in Philadelphia, which is called the International Solid State Circuits Conference. I borrowed this illustration from there, and it shows, on the vertical axis, the number of person-months required for the definition and design of the typical product of the time. And down below I'm showing the time starting from 1960 where it was almost immeasurable, and as you can see, in the late 1970's, we are dealing with something in excess of 300 man-months of design work.

To scale that a little bit, a very off-the-cuff estimate would be-of course, you have to consider at the same time that this phenomena is taking place, the cost of designers per person-month is also going up-the skills required are going up, there's inflation going on, so the actual cost is rising even faster than that. A typical cost of a typical product today-and that is not the state-of-the-art product-a typical product would be maybe in the 200 to 250 personmonths development cycle, maybe $1,000,000 expenditure. More extreme cases today-one example I think Mr. Sevin referred to here, he quoted $3,000,000-we have, unfortunately, run into $10,000,000 bills. This is just the design of the particular product, nothing to do with the actual processing sequence that you saw on those slides, or the complexity of the processes.

Mr. KASTENMEIER. Would you, or have you reached a certain plateau in that regard, limited by human resources in a sense, using your analogy to the jigsaw puzzle, whereupon it will take a computer in fact, something about human capability, to reach another level of difficulty or complexity in circuitry?

Mr. GROVE. We have not reached a ceiling in capability. What we have reached is an explosion of time, in terms of time, it takes to achieve each successive step. It is still possible to do. It just, like a jigsaw puzzle, takes an enormously larger amount of time in each new step of technology. As Mr. Sevin pointed out, we are desperate

ly trying to computerize as much of the steps as possible, but even though we have all kinds of computer aids at the beck and call of the designer, the process as a whole has not been-has not yielded to computerization. And I tend to share his view, which is a peculiar view for those of us who produce computers by the millions, this process will not be fully computerized, and therefore there's no bending of the curve, slowing down of that curve in sight.

The point I would like to make with this, is that this phenomenom of photographic copying of design, which we call chip pirating-actually, in all frankness, outside of occasions like this we call it worse, but we'll have to stick with that name for this occasionhas always been present in the industry, but it has not been either happening as frequently as it is today, nor has it been as painful. When the actual process, when we were down in the low level of the number of man-months required to produce a new product, the temptation which would put a reputable manufacturer, as you put it, into the practice of copying somebody else's design was substantially less than it is today. By the same token, the damage that he would cause to the originator of that design, was also relatively less. As this process is skyrocketing and going to the Moon, as that slide is illustrating, both the temptation is increasing, and the damage is increasing. And this is why we are coming to you and we are asking for assistance.

Just to illustrate out of my little bag, a couple-not an exclusive list-of this copying process, here is two examples that you have in the attachment. To the left is one of the most advanced products of our company. It is a 4,000-bit memory circuit that is a very, very fast memory circuit. It can work with very fast computers. It's a unique product. It is a product, in fact, that is so unique that the first time to my knowledge that IBM Corp. has ever gone on the outside to purchase full memory systems was in conjunction with our company's unique memory systems, using this component available, it was so unique. On the right, you have a photographic copy introduced within the last few months that we purchased on the open market, it's available for anyone to buy, an exact copy of that circuit by Toshiba-it's a Japanese electronic manufacturer. And not to imply that they are a commercial threat, but I have a very interesting discovery that we just ran across a few weeks ago: A photographic copying of one of our memory circuits, on the right this time-it's an old memory circuit, 5 years old-by Russians. This picture is taken-we do not have access to the chip, it comes from an article in "Aviation Week" that appeared approximately 1 month ago. So, this phenomenom knows no national boundaries, does not even know geopolitical boundaries, evidently.

That is all for the slides.

To recap the point, our industry thrives on the increasing complexity of our product. Increasing_complexity means enormous design risks and design expenses. Risk, because not all of them work; you pour these millions of dollars into it and they may or may not work. The damage to us when finally the effort bears fruit and the product works, and somebody can go and with essentially a photographic camera attached to a microscope and very little sophistication and can take advantage of what we have invested, is enormous. The consequence of that damage, if not arrested, is that

nobody, no responsible management can authorize this expenditure of millions of dollars of development effort if it is known in advance that the moment it is going to work and it's going to hit the marketplace, somebody is going to take your leadership away by the act of copying. I can give a person illustration of that. This morning, this very morning, I was involved in a review meeting of a major project, a project in which the design investment is well in excess of $10 million and the major issue overhanging whether we continue with the rate of investment or not, was searching our brains as to the likelihood that we will be able to protect those designs once they hit the marketplace from exactly this type of copying. We have continued with that project today, but if this pirating and the trend in increasing pirating continues, the day will come when the aggressiveness with which we are investing in product development will stop, and that will be the beginning of the end of leadership in semiconductor technology for our industry and for our country. That is why I am asking for this committee's support.

Mr. KASTENMEIER. Mr. Borovoy, would you care to make a statement?

Mr. BOROVOy. No.

Mr. KASTENMEIER. Mr. Grove, you would concede that if protection were possible for this area to the extent that you've requested for reasons you've recited, you do concede, do you not, that 75 years is not a necessary term of protection?

Mr. GROVE. It is not a necessary term at all.

Mr. KASTENMEIER. What would be a minimum necessary term of protection?

Mr. GROVE. Ten; seven to ten.

Mr. KASTENMEIER. Seven, ten, fifteen years, to be on the safe side?

Mr. GROVE. Yes.

Mr. KASTENMEIER. That's to be on the safe side. Actually, 7 years would probably be adequate, given the fast-moving nature of the present technology?

Mr. GROVE. That is exactly the point. In this industry, things have changed fast enough that 20 years-there is probably very, very few semiconductor products or integrated circuit products today, that are being produced today, that were made even 10 years ago. It is in the first several years of the existence of the product that the protection is necessary. The producer of the product has a lead, and it is that lead that needs to be protected. Mr. KASTENMEIER. Mr. Grove, or Mr. Sevin, what efforts have your companies made legally to protect, apparently unsuccessfully, but what alternatives have your legal departments considered in an effort to protect these designs to date?

Mr. BOROVOY. Let me take that one, if I could.

I'm Roger Borovoy, also from Intel Corp.

Obviously, most of these products that were spoken of today have one aspect or another covered by a patent. For example, the process of making it can be covered by a patent. Certain of the circuit elements which might cover a small part of it would be covered by patent. So that's one area. As has been explained, the patents are