Dueling Fools
Deep in the Heart
of Texas Instruments
January 13, 1999

Texas Instruments
Bear's Den
by Chris Rugaber ([email protected])

Want to invest in the fast-growing DSP chip industry and make some quick money? Oh, sorry, missed your chance: Texas Instruments has already jumped 81% from October 8th to December 31st. Maybe next time. Try to move a little more quickly, though, OK? I mean, hey, if you're going to play with the big boys, you better move! None of this namby-pamby research and due diligence stuff.

Just kidding, of course. It would be very unFoolish to write off a stock just because its price recently spiked upwards. And while TI's price increase is interesting, and demonstrates how quickly the market is jumping on any Internet-related stock, I'll look at that more in the rebuttal. There are plenty of other reasons to be wary of Texas Instruments.

I'll assume my bullish colleague has explained what DSP chips are and why TI is looking to them for salvation. Yes, the DSP market is expanding and they are pretty neat little thingies, and they're likely to be key components in the integrated voice/video/data networks of the future, as they already are in cell phones, DVD players, hard disk drives, and so forth. But this doesn't mean Texas Instruments has an easy ride in front of it. Quite the opposite, in fact. Don't make the mistake of assuming that just because a company's market is expanding, the company's cash flow and earnings will as well.

For example, while many people point to the growth of the Internet as a reason to invest in Cisco (as I did in my last Dueling Fools), Cisco has a track record of several years of stunning growth in sales and profits, and is doing business in a variety of innovative, efficient ways. It's growing with the Internet and is likely to continue doing so. Texas Instruments doesn't have any such track record.

After all, TI just began focusing on DSPs in the last year or so. Many of you may think of Texas Instruments as the people who made your calculator or your digital watch. The company also had a defense electronics business, an enterprise software business, and in 1995 it pulled in $13 billion in revenue thanks, in part, to its memory chip business. However, price collapses in DRAM chips clobbered TI as well as the rest of the industry. So after revenue declined to $9.75 billion in 1997, TI decided to make a few changes. As the company's website trumpets, it dumped 12 operations in 20 months and is now focused on becoming, as a TI press release from last September describes, a "pure-play digital signal processing� and analog company." Well, here's hoping that "pure-play" gamble works out!

OK, so it's not really a gamble. No one disputes that DSP chips have a pretty good future in front of them. In fact, everyone knows they do, and that's why everyone wants to start making them. This is one of TI's potential problems: competition.

In 1997, TI had 45% of the DSP chip market, to Lucent's <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: LU)") else Response.Write("(NYSE: LU)") end if %>28%, while Motorola <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: MOT)") else Response.Write("(NYSE: MOT)") end if %> and Analog Devices <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: ADI)") else Response.Write("(NYSE: ADI)") end if %> had about 12% each. Others are entering the market, such as IBM and several European firms, including Siemens and Philips. TI's current competitors are showing no signs of slowing down. Lucent and Motorola have launched a joint DSP effort, and one of their "StarCore" chips in late 1999 will be able to process 3,000 MIPS (millions of instructions per second), while TI's leading chip can handle only 1,600. On top of that, Analog Devices announced in October that its upcoming TigerSharc chip will be able to process 5,000 MIPS.

Aside from giving TI a run for its money, these developments demonstrate that DSP technology is constantly evolving, which makes it harder for investors to follow what's going on. A 5,000 MIPS DSP, for example, is really a "supercomputer on a chip," as an Analog Devices executive said in a November 30th Business Week article. There seems to be a consensus that DSP technology will soon be part of these broader-functioning chips, instead of standing alone. While TI has introduced products to keep up with this trend, an article in Electronic Buyers News reported that "As the DSP market moves to system-level integration, many nontraditional suppliers believe opportunities are opening for multiple approaches to DSP functionality." In other words, more competition!

Another reason to be careful with TI is that it has yet to really show a bang-up bottom line. Becoming the king of DSPs may not be a bad business strategy, but the verdict is still out on the execution of the strategy. In the third quarter of 1998, for example, its sales of DSPs only increased 17% from the year-ago period, which is a reflection of economic instability abroad and declining prices for DSP chips. (Sales had increased 37% in 1997 vs. 1996; predictions for the DSP industry overall are 30-40% a year).

Even after removing its money-losing memory business, 1998 3rd quarter revenues were 8 percent below the previous year's. Assuming that TI meets the First Call consensus estimate of 54 cents a share in the fourth quarter, this would give it 1998 EPS of $1.09 and a trailing P/E, at January 4th's closing price of $87.125, of almost 80. While significant growth in EPS is possible, such a high price for a company that's just finished completely overhauling its business and is refocusing on a fast-moving industry is excessive.