Magazine

Q&A: A Chip Veteran Gazes into the Post-PC Future


If the 1990s come to be known as the digital decade, Texas Instruments (TXN) CEO Tom Engibous won't be entirely content. For TI, the king of digital signal processors (DSPs), the recent past and near future are also about analog. It is the company's unwavering focus on chips that process natural waves--what the human eye sees or the ear hears--that has kept TI in the chip industry's vanguard.

In the era of mobile, networked communication, DSPs and analog devices matter more than ever, says Engibous, a 26-year TI veteran. These critical semiconductors process speech in cell phones, transport images in digital cameras, and stream Internet packets through DSL modems.

Indeed, the diversity of products that use TI's chips is one reason the Dallas company has survived the sector's withering boom-and-bust cycles, says Engibous. It should also give TI an edge in the post-PC future. Last year, global chip sales crumpled by $66 billion, down an unprecedented 32% from 2000's record numbers. TI mirrored the market: Its revenues in 2001 fell by 31%, to $8.2 billion, and the company posted a loss of $201 million. Thanks to its strengths in mobile and network technologies, however, TI's orders swelled 20% in the past quarter. And while PC sales continue to slump, prolonging the pain for TI rival Intel Corp. (INTC), Engibous foresees 10% revenue growth in the second quarter. Earlier this month, the CEO discussed the reasons for his optimism with a panel of BusinessWeek writers and editors, including Industries Editor Adam Aston. Here are some edited excerpts:

Q: How will TI fare in the post-PC universe?

A: There have been four eras in the electronics industry. Growth in the 1970s was driven by the mainframe computer. Your customer was a centralized department, so a chipmaker didn't deal with large numbers of buyers. In the minicomputer era, the number of customers went up, and the number of machines per person went up by 10 times or so. Then, in the '90s, came the PC era, and it moved toward one computer per person. That was the era of Intel and the microprocessor.

Now, we're in the next phase--call it the communications era--and diversity reigns. Instead of one platform, there's a proliferation. Instead of one computer per person, it's several computers--a phone, an organizer, or an MP3 player. It's an era of real-time signal processing. Players that are in that space--analog chips, DSPs, and other types of signal processing--are the champions of this era.

Q: What does this mean for growth?

A: You're going to see less of it this year. But over the long term, growth will be very good. Instead of one dominant platform, you've got many. The pattern of growth won't be as streaky, and it won't be tied to a Microsoft (MSFT) release or a new piece of hardware. We'll still see the cycles--this industry isn't capable of eliminating that--but they won't be as sharp as in the past.

Q: What applications will drive this?

A: When we look back on this decade, the two areas driving electronics will have been the mobile Internet and broadband communications to and throughout the home. Consumer media devices will become fully digitized audio, video, etc., all connected up to the Net.

Q: But the rollout of 3G--the future generation of mobile-phone technology--has all but failed even in Europe, let alone the U.S.

A: 3G never melted down, it never existed. The investments melted down. We still believe 3G will happen and it will be big. There are whispers in Europe that governments are going to pay back some [of the multibillion-dollar spectrum bids by cellular companies]. Europe is very proud of its wireless communications victory over the U.S. We shouldn't underestimate the fact that they will not let that victory go by the wayside.

Q: So when will we see meaningful, usable mobile Internet technologies?

A: If you look at the evolution of new technologies as complicated as 3G--which is more complex than anything to date--it takes longer than first thought. It took 10 years to go from analog cellular to digital cellular. As we speak, 3G is not technologically secure. You can't build a network and apply services to it. We found that out in Japan, where there are problems with 3G. Yet they will be solved, probably later this year. Also, the standard is extremely complicated, and people have interpreted it differently. That brings delays, but huge advantages to the first player with a successful infrastructure. In the GSM world, Nokia and Ericsson had such an advantage.

This gets to another challenge: services. The network operators today don't develop services, they buy them. They don't even do their own billing. The voice network wasn't put together by the operators at all--it was handed to them. So they're not good at providing services. And [with the mobile Internet], we're talking about services that have to inter-operate. Nobody tolerates a cell phone that won't connect with another network. Likewise, you won't be happy if you can't access multimedia services provided by a different operator.

Q: What about in the U.S., where a lack of standards has slowed cellular services?

A: We're disappointingly behind on all aspects of the spectrum. There are financial issues here, too, as in Europe. Someone has to pay for network development. And nobody seems to have that answer yet.

Q: So if the mobile Internet is still a ways off, where will near-term growth come from?

A: The growth is predominantly driven by the fact that our customers--the people building boxes--have whittled through their mountain of inventory. The [original equipment makers], the contract manufacturers, and the distributors, they're all still decreasing inventory. But now, they're running out of selective parts, and they're having to order those again.

Q: Does that mean fresh demand for tech gear?

A: It won't be what we had in the late '90s, when growth was 20% or more. Now, it's single-digit. Notebook computers are growing. And cell phones will still grow about 5% this year. Digital consumer devices--DVDs, read-write devices like recordable CDs, and digital still cameras--are growing very nicely. The wireless LAN is on path to grow from 8 million to 15 million units. The number of DSL lines will increase this year, too. You just aren't going to see explosive growth again until we see broadband deployment and the next generation of data services in the wireless sector.

Q: After the trauma of the past year, has much changed among the top chip players?

A: We're really still in the back end of the worst downturn in the history of the semiconductor industry. And this just so happened to coincide with a major new technology investment turn: a brand new lithography of 130 nanometers, a move to new copper-based materials, both coupled with a shift in the diameter [of the wafer from which single chips are cut] to 300 millimeters. The net result is that a significant number of traditional semiconductor strongholds have opted out of continuing on that treadmill. In fact, there are only five companies that have made the commitment: Intel, IBM (IBM), TI, and, in Taiwan, UMC (UMC) and TSMC (TSM). The Japanese have opted out--perhaps not consciously, but the result is the same. Nobody in Europe is investing in that. Even the Koreans haven't yet made the commitment. There's bifurcation of the haves and have-nots of technology.

Q: What about Japan?

A: The Japanese semiconductor industry is really hurting today. Whenever you read about alliances in our industry, that's a problem sign. Companies figure if they can't keep up, get married. Yet there isn't any evidence, period, in our history of where that has generated some type of positive benefit.

Q: Once you separate the haves from the have-nots, how does the competition shape up?

A: The contracts to do high-volume consumer applications will go to the players that have the ability to develop and build their own technology and to do large system-on-a-chip integration. Today that's a limited group. Of course, when the upturn comes, there will be players that come later. For now, our largest competitors have opted out. They've decided to source that leading-edge technology from somebody else.

Q: What about competition among the haves--those that have the latest chip technology?

A: It's interesting. If you look at the companies that design and make integrated circuits today, it's TI, IBM, and Intel. None competes [with one another]. Intel is in one space, TI is in one, and IBM is in another. I'm sure there's going to be some overlap, but not right now.


The Good Business Issue
LIMITED-TIME OFFER SUBSCRIBE NOW
 
blog comments powered by Disqus