ANDY GROVE HOW INTEL MAKES SPENDING PAY OFF
(FORTUNE Magazine) – Intel, a standout stock market performer of late, looked like an also-ran just seven years ago. Since then the company has transformed itself from a money- losing producer of commodity memory chips into the world's largest semiconductor company -- and the business best positioned to profit from the enormous increase in personal computer sales. Technology was only one element in the transformation. Just as vital was the willingness of CEO Andrew Grove to bet billions on research and development and on new manufacturing plants. In 1993 the company will put $2.5 billion -- equal to 43% of last year's revenues -- into R&D and capital outlays. Price wars are creating a mass market for PCs, which now outsell TVs and VCRs despite typical unit prices of $2,000 or more. Almost all PCs employ at least one Intel chip, and the company has a near monopoly on the leading-edge microprocessors -- most recently the 486 -- that do their thinking. Once investors connected the PC sales surge with Intel's market dominance, they scrambled to buy the stock. Intel shares rose nearly 30% from New Year's to late January. Grove, 56, fled his native Hungary in 1956 with $20 in his pocket. He earned < a Ph.D. in chemical engineering at Berkeley and in 1968 helped found Intel. As the stock was soaring, Grove met with FORTUNE's Stratford Sherman at Intel's spartan headquarters in Santa Clara, California. During six hours of interviews, Grove explained the underpinnings of Intel's success and shared his best ideas about management. Excitable and quick to challenge others' views, he contends that the lessons from his company's experience apply to any business that must adapt to a fast-changing marketplace. Excerpts:
THE STOCK PRICE -- Our stock has gone from $47 to over $110 since last summer. When you are in a competition and somebody raises the stakes, you've got to put out more. To meet the raised expectations of investors, we've got to go into an even higher gear.
THE NEW COMPUTER INDUSTRY -- The breakup of the old computer industry is what gave Intel its chance and made the mass-produced computer possible. The old computer industry was vertically aligned: Each company sold a completely integrated product based on its own proprietary technology. Companies like IBM and Digital Equipment designed and built their computers from the bottom up -- silicon chips, software, disk drives, everything. These vertically integrated companies would compete against other vertically integrated companies, and buyers had to commit to the whole package of one manufacturer or another.
As my sketch shows, a new horizontal industry model is replacing the old vertical one. In the PC age, everybody's products have to work with everybody else's products or they don't sell. Intel's processing chips will work in any IBM-type personal computer, no matter who builds it. The operating software that enables the PC to work is also standard. Application programs are sold in stores, like CDs or books, and when you install one, you expect it to work. Businesses compete for market share within each horizontal specialty. Intel is up against companies like Motorola in the basic silicon architecture. In operating systems it's Microsoft vs. Apple vs. IBM vs. Novell. Borland competes with Lotus and scores of other companies in all the varieties of applications software, like spreadsheets and word processors. And then you've got the Compaqs and Dells and many others making and selling the PCs. The old competition based on the advantages of one proprietary technology over another has been replaced by one in which you win by providing basic value. The new computing industry may be the best example of what I call an industrial democracy. It resists central guidance. Nobody can tell anyone else what to do. Your PC might have a processor from Intel, a display from Sharp, a hard disk from Conner, memory from Toshiba, a modem from U.S. Robotics, an operating system from Microsoft, applications from four different vendors, and yet it all works together. If it didn't, none of these products would sell. Industrial democracy is brutally competitive, and it can get very messy, but it enables technology to move very fast. Take skis. Any ski boot works with any binding. Any binding fits any ski. That permits innovation to take place independently in boots, bindings, and skis.
SPEED -- These changes in industry structure aren't happening only in PCs, they're happening everywhere. In any industry, as you go from vertical to horizontal, the customer isn't locked in anymore. And when companies lose their proprietary advantages, speed seems to be what matters most. In all businesses, to provide value, you've got to do whatever it is you do fast and with immense efficiency. Why do you think everybody is buying computers? It's not that they render you more productive, in the sense of more physical output per hour. They permit you to make decisions and produce results faster. The whole leveraged-buyout wave of the 1980s was made possible by Lotus 1-2-3 running on PCs. The necessary analysis would have been impossible without computers. In my own little corner of the universe, a lot of other people can build microprocessors. If we hadn't kept moving the technology rapidly -- from the 8088 chip in the original PC with 30,000 transistors, to the 486 with 1.2 million transistors, to the forthcoming Pentium chip with over three million -- we wouldn't have an advantage. Already a competitor has taken over 50% of the market for 386s. We're focusing on the 486, which no other company is selling yet, and the Pentium chip, which we plan to release this quarter. And we're already beginning to invest hundreds of millions of dollars in the generations that will follow. The pace of work these days isn't easy to live with, but welcome to the Nineties. Intel didn't create this world, we're just supplying the tools with which we can all work ourselves to death. Exhausting as it is, it's highly preferable to being unemployed.
INVESTMENT , -- You can't hesitate or hedge your bets. You can have the best product in the world, but if you fail to invest in enough plant and equipment to satisfy demand for it, all your efforts are wasted. All you've done is create an opportunity for someone else. In this business you have to build your own capacity -- that much vertical integration is essential. A leading-edge product requires leading-edge manufacturing capability, and you can't buy it. What's available from others is usually yesterday's technology. You can't build a winning Formula One race car out of stock parts. The investments Intel makes are enormous, and they don't always pay off. In the past six months we have killed three major projects into which our engineers had sunk a total of maybe 200 to 300 man-years, at a cost of $30 million to $40 million. In the case of the Pentium chip, we have to invest in substantially more plant capacity than we needed for the 486 generation. Not only is the whole PC market growing, but the chip itself is bigger than the 486, so we get fewer usable chips out of each wafer of silicon. To keep our lead we absolutely had to go on this very large capital spending program.
THE INFURIATING PC -- The old mainframe computer companies were masters at integrating all the necessary pieces into a seamless system. In the PC world, we're all on our own. Each company in each horizontal layer designs its products to work with everybody else's products just well enough to sell. We have to have a minimum of integration to survive, but doing more is a luxury. That's why these products -- chips and programs and computers -- are not integrated enough to be easy for people to use. It took me 40 hours to get my first laptop going. At Intel we have put the first ''mission critical'' application on our PC network: a companywide electronic mail system connecting 20,000 employees from Hong Kong to Moscow. Two years ago we hardly used E-mail. Now we'd be paralyzed without it. It was an enormous and very complicated job to put this thing together. Then, as the traffic grew to 300,000 messages per day, the system got clogged up. I hear the Clinton Administration wants to create a unified E-mail system for the whole federal government. That's hilarious. We are not particularly incompetent at Intel, and we've been working on our little E-mail system for two years. Saddam Hussein will have free rein while the government tries to get its system to work.
VERTICAL COMPANIES -- I wouldn't categorically say that companies structured the old way can't survive, but it's hard to see them thriving. Anything that can be done in the vertical way can be done more cheaply by collections of specialist companies organized horizontally. But there's a huge business opportunity for the old mainframe and minicomputer companies. Their very structure forced them to develop expertise at integrating layer after layer of technology into large-scale systems that work. The demand for that expertise is enormous, because the horizontal industry arrangement obviously doesn't satisfactorily resolve the issue of systems integration. Networking is changing the way people use PCs. Intel's troubles with E-mail are typical of the difficulties users face when they try to link large numbers of PCs into networks. Nevertheless, the growing need to communicate quickly by computer is making such networks a competitive necessity for more companies. Advances in communications software will increase the demand for large networks, and thus for capable systems integrators. I wish people would stop complaining about lousy integration and get into this business instead. They could call themselves Mission Critical Applications R Us. It's a fantastic business opportunity. The only thing preventing the old-line companies from exploiting this is mind-set. Systems integration is a service, and these companies grew up bending sheet metal. They like bending sheet metal. These companies should go where the business is. They ought to make this move while the going is good.
TRANSFORMATION -- A corporation is a living organism, and it has to continue to shed its skin. Methods have to change. Focus has to change. Values have to change. The sum total of those changes is transformation. One of the biggest lessons I've learned is that it is always easier to put strategic changes into action than to declare them as policy. Ten years ago Intel's focus was largely on semiconductor memories, particularly a type called DRAMs. In certain respects this was a very simple business. If a 4K chip was good, guess what? An 8K chip was better, and 16K better still. We had only a few customers, so with a small group of salesmen we could cover the whole U.S. Today we're a microprocessor company, with 100 million customers. Not the people who pay our invoices, but the end users of PCs, whose minds we have to / win in order to succeed. We have to figure out what their needs will be years from now, what sort of application software they'll be using. Then we have to design microprocessors that meet those specific needs. We're not pure technologists anymore. We now must work with companies on all the other levels of the PC industry. And to reach end users, we're running TV commercials. Add up all the changes, and we're a different company than we were.
Look at this sketch of the transformation curve. There is at least one point in the history of any company when you have to change dramatically to rise to the next performance level. Miss the moment and you start to decline. For us the turning point was somewhere around 1985, when we decided to get completely out of DRAMs, the product our company was created to make. Our rate of change got accelerated by the fact that we were losing a lot of money. Emotionally, it's easier to change when you're hemorrhaging. We came to realize that the formulation and articulation of strategy is the job of top management. Back then we had the very ridiculous system, common in American business at that time, of delegating strategic planning to strategic planners. The strategies these middle managers prepared had no bearing on anything we actually did. While they were writing reports and making presentations, our production planners were already doing what the business required: shifting our capacity from the DRAM business that the Japanese were taking away from us to this new business of microprocessors, 286s at that time. But until top management recognized that change and articulated it as a strategy, Intel suffered from what I call strategic dissonance. While we were shifting capacity to microprocessors, we still had our best development people working on memory projects. That made no sense. So a major change in strategy can only happen when management reconciles action and speech. Denial can blind people. We found it very, very hard to face up to the DRAM decision. Financially, it should have been easy. In retrospect, it should have been easy strategically too. Yet we, Intel management, were at each other's throats over this. Our people have navigated successfully through one transformation, so perhaps it won't be as hard to sign them up for another one. But success can trap you. The more successful we are as a microprocessor company, the more difficult it will be to become something else. To take advantage of some of the opportunities I see ahead, we're going to have to transform ourselves again. The time to do it is while our core business is still so strong.
THE SOFTWARE SPIRAL -- The PC marketplace is driven by an invisible momentum that I call the software spiral. The most advanced software usually runs sluggishly on existing hardware because the new programs require the microprocessor to do more complex work. So each generation of software makes users anxious to move up to the next generation of chips. Microsoft Windows was a classic example. It was supposed to work on a 286-based PC, but you really need at least a 386 to quickly process all the data that go into those graphic images on your screen -- and even a 386 isn't totally satisfactory. We believe that the new software applications of the coming years will make users crave Pentium-based PCs.
DIGITAL MEETINGS -- The next big category of applications that I foresee doesn't even have a name yet -- call it electronic conferencing. It's the next step beyond E-mail. Electronic meetings will take place in real time and include sound and video and the ability for people to work on documents together. I hate traveling. People in business travel constantly. It's uncomfortable, it's expensive, but most of all it slows you down. Think how much time we could save if we didn't have to travel, and how quickly we could transact business! To avoid plane trips, Intel has become a big user of video conferencing, but it fries me that in 1993 we still have to fax a piece of paper so that everyone can look at the same document. We don't spend money on video conferencing to smile at each other and say ''Hi.'' We do it to discuss some kind of business issue, so looking down at a piece of paper with information on it is important. We already have technology capable of sucking all those capabilities -- video, audio, E-mail, and shared access to documents -- into a PC. We're betting that some portion of our 100 million customers will kill for electronic-meeting capability. Intel will end up spending $2 billion or $3 billion on R&D and plant capacity for the Pentium chip, which we designed to be very good at processing video images. (Our total investment in the 486 product line is closer to $1 billion.) We'll know whether we made the right bet in a couple of years.
STAYING ON TOP -- Maybe we can -- but it won't be like falling off a log. Our corporate mission is to be the preeminent supplier of building blocks to the new computing industry. If we develop the right building blocks, we'll win. If we are wrong, we'll fail. There's no competitor around who can do as much damage to us as we can do to ourselves. Technology can make you lose, but it can't make you a winner. First, you have to start with a better product than the other guy's. Second, it has to be better in a way that your customers can appreciate. And third, your customers have to know it's a better product in a way they can appreciate. Intel's transformation has to continue. We've got to become more than a microprocessor company. We see three distinct ways to market our building blocks. One is selling the basic microprocessor. The second is selling other types of chips -- those that control the disk drives, or process video, or do other auxiliary tasks in or around the computer -- to PC makers. The third way is where I believe our future may lie. It consists of selling products to individual users that make your computer more capable, such as fax modems or network boards. Looking ahead to the world of electronic meetings, we see big opportunities both in powerful processors and in other products that turn the world's 100-million-plus PCs into multifaceted communication devices. We'll keep on investing in the capacity to produce our leading-edge chips in quantity. As those chips become available in large volumes, we'll do some not- so-gentle nudging on price to encourage people to migrate on to the next generation.
INTEL'S CRITICS -- We have gone to the courts to protect our intellectual property from competitors who would have liked to help themselves to it. I understand that you don't make friends that way. But we spend tons of money on R&D. As a company with a disproportionate amount of intellectual property, we are forthright about stating that our property is valuable, and we want to prevent other people from helping themselves to it. As for our relations with PC makers, the main issue is that in many instances we are a unique source of technology for our customers. When we don't do a flawless job of supplying the product they need, we make people unhappy. Intel sells a category of chips called flash memory, which store lots of data at relatively low cost. We created the market and stimulated demand, but our output has been very disappointing. We're six months to a year behind on our orders. I'm not proud of that. As long as we send one customer away unsatisfied, I can't take full pleasure in our performance.
CHART: NOT AVAILABLE CREDIT: NO CREDIT CAPTION: INTEL