The brainmaker.

Will LSI Logic's "System-On-A-Chip" become the electronic brains for a whole new generation of Internet appliances? Just ask any 10-year-old who owns a Sony PlayStation.

"The Internet is the computer," has been the mantra of Sun Microsystem's Scott McNealy since 1989. Six months ago, Oracle's Larry Ellison announced his vision of what he calls a "network computer" - a $500 open-systems machine that, when linked to the Internet or corporate network, could substitute for a more expensive PC. Are the PC's days numbered? "Macrocosta" author George Gilder thinks they are. Intel CEO Andy Grove dismisses the Internet PC as the Cabbage Patch Doll of computing. Even if one believes that people like Gilder and Ellison hyperventilate about this, it would be unwise to discount the appeal for low-cost alternatives to the standard desktop PC. After all, the typical office spends $40,000 over five years on a single PC when software and maintenance are factored in, according to The Gartner Group, a technology consulting firm.

Unlike an ordinary PC, which requires complex operating systems such as Windows 95 and runs programs on its hard drive, a simpler teleputer would download an operating system from a network and store data on a remote server. The Financial Times reported that Boeing has expressed interest in buying 100,000 of these so-called "cheap, dumb" terminals.

Among the first to unveil the technology that would make this controversial computer concept - among other products - a reality is Milpitas, CA-based LSI Logic, a $1.3 billion ASIC chip maker. ASICs, or application-specific integrated circuits, are customized chips designed for highly complex and specific functions. By taking the lead in integrating many functions on a single chip, LSI has leveraged its expertise in gate arrays (a particular type of ASIC) to come up with its "System-On-A-Chip," which enables such low-cost devices such as Sony's PlayStation or Thomson Electronics's set-top boxes used for digital satellite broadcast.

LSI - an acronym for "large-scale integration" - was founded in 1981 by Wilfred J. Corrigan, 58, former CEO of Fairchild Semiconductor, whose previous experience working for Motorola earned him the distinction as the godfather of Silicon Valley. The son of a Liverpool dock-worker and trained as a chemical engineer, Corrigan took LSI public in 1983. Thanks to its contracts with the military and aerospace customers, the company thrived during the 1980s, when Japanese memory chips flooded the market and threatened many U.S. chipmakers. But by 1990, a year after Corrigan had undergone triple bypass surgery and almost lost the company, LSI faced a serious crisis in the decline of mainframes, which used a lot of custom circuitry. In 1992, the company laid off a few hundred employees and shut down an obsolete German factory. Corrigan assembled his lieutenants - including EVP Brian Halla, formerly of Intel; Cy Hannon, EVP of worldwide operations; and Moshe Gavrielov, SVP of international marketing and sales - and called on them to rethink the company's future.

The team shifted to more complex projects that used its expertise in miniaturization - allowing customers to create entire systems on a single silicon device that formerly required dozens of separate chips. The CoreWare process, which allows up to 5 million gates to be embedded in a chip, is the heart of LSI's architecture. Its G10 technology can put up to 49 million transistors on one custom designed chip. By contrast, Intel's Pentium Pro, the industry's new PC standard, holds 5.5 million transistors on a single chip.

Whatever one assumes about the growth and character of the Internet, these chips, which cost LSI's customers about $35, can be optimized for a variety of uses, including video, such as the digital video disk players that will be released later this year. They could readily be used for a variety of set-top boxes for integrated entertainment systems incorporating video, disk, and CD-ROM. Corrigan believes that the first of these so-called "Internet appliances" may be introduced by the end of this year.

Corrigan's challenge, however, was made more difficult in early May when Halla, his top lieutenant, presumed successor, and head of new-product development, resigned to take the top job at National Semiconductor. Corrigan reacted quickly to promote Gavrielov to EVP of LSI's products group, and soft-pedaled the impact of Halla's departure. In a prepared statement for CE, Corrigan said, "In Silicon Valley, it's commonplace for executives to switch jobs and climb the corporate ladder. Brian played an important role in articulating LSI's product strategies, which are now being executed and implemented by a team of talented executives and managers."

CEs J.P. Donlon recently caught up with Corrigan at LSI's headquarters near San Jose.

- J.P. Donlon

In your view of the world, what will give us access to the Internet, and how will these products fit into the market?

The PC is pervasive, but the telephone, TV, and even video games are even more so, and any one of those might be an Internet access device. At LSI Logic, we don't want to be in the Internet box business; we just enable others to be in the business. I think there will be Intranet boxes and Internet boxes. There will be $500 boxes and $80 boxes, aimed at different parts of the market, which will hook into your TV. For the last 20 years, people have been carefully trained at how to flip channels - so at one extreme there might be a low-priced, sub-$100 access device for the couch potato.

Meanwhile, corporations will use more sophisticated Internet access devices. Today, everyone in an office gets a customized PC, and every secretary out there has a slightly different PC hanging onto the network. But look at telephones: They're all the same. They plug in the same way, and they aren't customized with different software. Internet devices will be like that: Versatility will come from the network, not the box.

So you are betting that the Intel/Microsoft model, in which the PC does everything, will be replaced?

Let's say you've just laid out $3,000 for a PC, and do a variety of things on it: Use it as a typewriter, use it for your personal finances. But what if little Joey says, "I want to run my video games on this"? The average parent will say, "This isn't a good idea." Joey's been playing video games on it and suddenly the family finances have all gone away. So the parent decides to just pay $200 and give him his own device.

There will be a big market for a standard PC. The question is how far can you drive the cost down by sheer volume with a PC in which it doesn't matter that you've got all this excess function. Our bet is that we can drive the cost down with a specific design.

Your vision of the future is that people will link up with Internet, or whatever the Superhighway becomes, without being PC-based?

Yes. Look at set-top boxes. You might have one for your satellite, one for your cable, one for your digital video disk, and one for your Internet access device. How many of these are you going to have on your TV? It's going to get very cluttered. The solution is to combine them all in one chip. You might still have a box - or maybe it's embedded in the TV itself.

And you will make that chip.

Yes. All these different functions can be shared by one chip. And it will be just a few things on the outside that will enable you to access particular functions. Do you want to access the satellite? Do you want to access the cable? Do you want to play the digital video disk? All the circuitry is exercised exactly the same each time.

So everyone will have one device?

I doubt there will be one God-awful device you can carry in your pocket that does everything. The Internet has changed a lot of thinking, because if you have an access device and you have a convenient plug, you needn't carry all these ganglia with you. So there might be a utility on a plane that you can just plug into.

So I think we're going to end up with many different, small devices that will be available at impulse-buy prices - a couple of hundred bucks.

THE BILL AND ANDY SHOW

Andy Grove calls the Internet PC a boyish dream. One of you is wrong. Will the Internet PC replace the PC?

Intel sees the world as PC-centered - with the PC as the sun of the electronics business and everything else rotating around it. I don't see it that way. Rather, I believe the interconnection is the center, and everything else hangs on. The PC is just one more thing that hangs on to the Internet.

If I were in Andy Grove's shoes, I'd have the same reaction. Intel has become No. 1 by having a single-minded focus. In 1985, it embarked on what has been called the "Tall Thin Man" strategy - dominating microprocessors - while the big companies then were practicing the philosophy that one company can make everything. Intel's proven to be right over the last 10 years. And none of the computer companies - IBM, Digital Equipment - could have predicted what would happen to the computer business, but Intel, meanwhile, could change course at each turn of the PC business, and with Microsoft completely dominating the architecture.

Bill Gates has said he does not believe in the Internet PC, that it's simply a browser device that anyone can knock off. Is he wrong?

I can understand why Bill Gates would not think it's a good idea, since all these PCs now require a Microsoft license. With an Internet PC, you don't need a license. You get into the Internet, you pull off whatever you need, and many of these basic programs are going to be basically free - you can get them from the Internet.

Does LSI Logic represent a threat to Microsoft or Intel, even in a small way?

I don't think so. What we do is only a portion of what happens. They will change their strategies to fit in the new equation.

If what your company says is true, the world will cease to be Windows/Intel centered. And that's a big difference from what we have today.

But the average person has a lot of centers in his or her life. There's the work center and the home center. And in those, there are many things, such as the TV and the telephone, that could get more intelligent. Look at the Sony PlayStation. In reality, it has a video disk player built into it, so with a minor modification it could be a playback device for video disks. So you could use it as a replacement for a VCR if there were an adequate supply of movies on disks. So you've got this video game thing that's also a disk player, and that could also be an Internet access device. You could put a keyboard on it and maybe use it as a word processor.

So just as you can play video games on a PC, you'll also have a video game device on which you can do some PC functions. Intel and Microsoft haven't figured that out yet. The eventual engagement with a large number of devices might be very profitable. If their position is, "If we don't dominate it, we don't want it," that will be a mistake, because the world will roll over them.

Will Microsoft and Intel make that mistake?

Long term, no. They'll have to fight the rear guard until it's definitely not going to win, and then they'll flip.

CHIPS 'R' US

What differentiates LSI Logic from its competitors?

We're the first company to reduce computer design to practice on a large scale - our engineering goes toward optimizing that process. We focus on methodology: How you crank out a design quickly accurately and at low cost. It's rather like Henry Ford. He didn't invent the automobile, he invented how to manufacture the automobile at high volume.

Of course, we've been copied in high volume. But we do more designs than anyone else, we've got a bigger design factory, and we've got more people working on optimizing design methodology. We're focused on the process technology for our particular application, which is slightly different than the process technology for memory, or the process technology for microprocessing. Another differentiator is our worldwide design infrastructure: We have 50 design centers around the world, and we can design locally - whether that's in Stuttgart, Osaka, Paris, or Minneapolis - whereas many of our competitors have to bring it back to the core, which limits the number of designs you can do.

What will set you apart in the next five or 10 years?

More of the same. More intellectual property embedded in the chip, richer libraries of elements for people to design with, more applications engineering support for the customers, and denser technology.

What factors will most influence your industry in the next three to five years?

The computer market is rapidly becoming just one leg of the stool, and communications and consumer products are becoming as important.

Communications 15 to 20 years ago simply meant transmission, and that was dominated by AT&T. Today, it means much more than that. Now, we classify cable companies as part of the communications business. The transmission part of it is actually going quite slowly, but there is tremendous growth in cellular and network. The Internet will accelerate that even more. So we're being focused more and more on end markets and end customers.

RESUSCITATION REQUIRED

LSI Logic went through a near-death experience in the late 1980s. How has the company been made stronger as a result?

When we went public in 1983, the reason people paid so much for our stock was our customer list. We had major contracts with IBM, Digital Equipment, Prime Computer, Norse Data in Europe. We supplied just about every minicomputer company and virtually all the mainframe companies.

In the late '80s, down went the Berlin Wall and with it the pillar of our business, the military business. Second, companies such as IBM and Digital Equipment got into deep trouble, and all the programs we designed never made it into full-volume production. Major customers such as Wang, Prime - half of Route 128 - died. We were stuck asking ourselves how we could have been so stupid that we didn't realize that IBM and Digital Equipment were going to get into severe trouble, and the whole minicomputer industry was going to go away worldwide. And we were stuck with a lot of capacity. I look back and wonder how I could have been so stupid, but then I say, "Well, just like all the guys who were [at the time] still investing in Digital Equipment at $150 a share." And IBM - God knows what the share price was at the time. No one really saw very clearly what was going to happen as the PC chewed into the bigger computers.

So in 1990, we were going to be a $1 billion company, but as these companies went through heart-wrenching episodes, we shared them. It took us a couple of years, but we realized we needed to look at different markets and customers. We restructured, but the fundamental thread of the technology is the same.

Describe the restructuring process.

It was like breaking tablets that had come down from Moses. The company had been founded on concepts we started in 1980, and they had to be changed. The toughest part was changing the culture, and people left because they felt they were being disloyal to the culture we built up in the 1980s.

Also, it wasn't that simple from my end. It's not like this wonderful plan just comes out of your head. We had a couple of stabs at it over about two years, trimming a little here and there. That wasn't enough. Finally, we had to rethink everything in 1992, since Wall Street was losing interest in us. We could have stumbled along as a modestly performing company that was well respected by its competitors and its customers but got no respect from Wall Street.

Speaking of Wall Street, your stock has fallen from $62 in September 1995 to about $33 now. With all these new developments - process technology, the fact that we're on the verge of STBs, digital video disks, Internet access devices - one would think Wall Street would be pushing the stock further rather than kicking it in the legs.

Wall Street has gone sour on the whole sector, but that's changing. We were also nailed by the big institutions that got out of technology in general, not just our stock. But if you look at the long-term cycle of growth rate, the stock's going to multiply. Compared with most of my brethren in the business, our stock has moved more or less in sync with the whole market in the last six months.