Betting the farm
(http://www.eb-mag.com/registrd/issues/9904/0499ltx.htm)
LTX's Roger Blethen has
bet everything on the
company's Fusion
single-platform tester for the
system-on-a-chip market. Is
it the right product at the
right time?
By Howard Rudnitsky
photograph by Stuart Rosner
Roger Blethen, CEO of semiconductor test equipment maker LTX Corp., is
possessed by a vision: the potentially dramatic growth of system-on-a-chip
(SOC). He fervently believes semiconductor manufacturers have all the
front-end equipment and most of the software they need to do the job.
Unfortunately, argues 47-year-old Blethen, the current generation of
multiple-platform, automated-testing machines can't adequately test these
increasingly complex new chips.
His proposed solution: a single-platform SOC tester, LTX's Fusion product.
After a year and a half in development, Westwood, MA-based LTX has
recently started to deliver its first Fusion testers to customers such as
National Semiconductor, Philips and Hitachi. In early January, the
company had orders for two dozen more.
SOC is a market that appears poised for rapid growth over the next few
years, according to Dataquest. The value of the market in 1998 was $9.1
billion, according to Bryan Lewis of Dataquest. He projects that by 2002,
the market will grow to almost $24 billion. If it does, that's faster growth
than the semiconductor market as a whole.
Warning signals
Unlike larger automated-test-equipment companies such as Teradyne,
Advantest, Schlumberger Technology and Hewlett-Packard that are
diversified in sectors such as memory, logic and mixed signal, LTX, with
$196 million in fiscal 1998 revenue, is tied largely to the mixed-signal
testing market.
However, by the end of 1998, LTX
saw its mixed-signal market share
decline to around 17% from 21% in
1996. Under pressure from industry
leader Teradyne's highly successful
Catalyst mixed-signal-tester product
line, the company shifted to the
expensive development of its
technologically advanced
single-platform tester. Over a
two-year period, Boston-based
Teradyne Inc.'s multi-platform
Catalyst share rose from around 39%
to more than 45% of the estimated
$900 million mixed-signal market.
LTX was hurt financially by the costly
closing of its San Jose
digital-manufacturing operation. It
then had to shift the duties of the California facility to its Massachusetts
plant at the same time it was developing its SOC tester. To make matters
worse, the semiconductor industry in 1998 encountered its worst slump in
years. Orders for testers fell sharply. As a result, LTX recorded a $78
million loss in fiscal 1998.
While all semiconductor equipment stocks were hit, LTX shares were really
hammered. By October they had fallen to just under $1 from a high of
around $14 three years earlier.
Investors wondered if the company could come out of its tailspin. It had
delivered only one of its new Fusion testers by the end of September 1998.
Meanwhile, its $69 million in cash had been reduced to $20 million in 15
months. There was even an unsubstantiated rumor that LTX might be a
takeover candidate. The market value of the company at its low point fell
to a mere $35 million.
Despite that low value, however, it still wouldn't
be very easy to succeed at a hostile takeover. LTX
has a staggered, three-year election of its board
of directors, it's domiciled in Massachusetts, which
favors in-state companies, and it has a "poison pill"
in place, which would make it much more costly by
issuing lots of new shares if a hostile takeover
were attempted. In addition, it's quite likely that
key technical people would leave the company.
Clearly, Blethen has had his hands full ever since
September 1996, when he won out over his rival, Martin Francis, to become
CEO. LTX's long-time Chairman and CEO, Graham Miller, had appointed
both men co-presidents in 1994 to see who would succeed him. Miller, 67, is
now chairman emeritus and still owns stock in LTX, primarily in the form of
stock options.
Since the low point last year, the picture has brightened somewhat for the
semiconductor industry. The ratio of the dollar amount of new orders
compared to revenue, the so-called "book-to-bill ratio," for the industry
recorded a four-month upturn. Not surprisingly, Teradyne's stock has also
recovered and hit an all-time high of $65 in February, then fell to $52
recently. LTX shares rose as well, recently trading at around $5.37 a
share.
According to Blethen--who graduated in 1974 from Boston's Northeastern
University as an electrical engineer and who worked for Teradyne before
leaving with a group of employees to help found LTX in 1976--that doesn't
mean it will be business as usual. He points out that semiconductor test
equipment orders for the industry through October were down by some
two-thirds not simply because business had been lousy for semiconductor
companies.
Blethen's mantra
Says Blethen, in what sounds like a mantra:
"The answer is that the new equipment our
customers require is only just emerging.
There's a major change going on in the
semiconductor industry. That's not just me
talking." He cites top executives at Lucent
Technologies, National Semiconductor,
Motorola and Texas Instruments,
contending that a major shift to SOC is
starting and its growth will be explosive.
Blethen argues that the industry's model is no longer how much smaller can
you make the chips. What's emerging is an emphasis on how much intellectual
property you control. All the new digital and mixed-signal designs now are
headed toward highly integrated combinations of digital logic, analog and
mixed signal and a little embedded memory. Now, says Blethen "marketing,
design and test are driven by what the chips do. You have to know what the
chip is going to do. So LTX has moved out of the realm of tester hardware
into the realm of testing intellectual property. With our SOC tester, LTX
has got at least a two-year lead on everyone else. It's been tough and
painful for us to get that lead, but we've got it now."
LTX's new Fusion testers can cost a little under $500,000 for just
mixed-signal machines to over $3 million for full SOC designs with more
pins and more mixed-signal content. The more complex machines provide
SOC solutions with digital VLSI, embedded memory and mixed-signal test
capability all in one platform. LTX has delivered several single-platform
Fusion testers and has many more on order.
For the time being, Teradyne is choosing not to take the single-platform
route but to continue to improve its highly successful multiple-platform
Catalyst system tester line. That approach seems to be paying dividends.
On January 19, Teradyne reported that its sales and earnings for 1998 had
lived up to expectations: revenue of $1.49 billion compared to $1.27 billion
in 1997; earnings of $102.1 million versus $127.6 million in 1997.
However, the reason its shares
jumped by nearly 10% that day
was that Teradyne's orders for
the fourth quarter were up by
50% over the third quarter, to
$371 million. Teradyne's
memory tester orders were up
the most, according to Tom
Newman, vice president of
corporate relations, "because they had been down to near zero in the prior
quarter. We also had a strong logic test business and the mainstay of the
semiconductor test business in the economic downturn--mixed signal--that
also grew a healthy amount in the quarter."
According to LTX's Blethen, semiconductor companies were supposed to be
holding back orders for mixed-signal testers because they had been
concerned that the existing capabilities of multiple-platform testers won't
be able to adequately handle the more demanding job of testing SOCs.
New orders mean what?
If so, how come semiconductor makers are now increasing orders for
Teradyne's multiple-platform testers? "Teradyne's fourth quarter orders
confirm the start of the next cycle," says Blethen. "[But] who will win the
race?"
Teradyne still isn't listening to the future needs of its customers, Blethen
says. A number of semiconductor makers are still reluctant to make major
tester investments, which can easily run well over $100 million a year, when
they are concerned that three years later, SOC increasingly may be in
demand and their existing line of multiple-platform testers won't be
effective enough, forcing them to turn to SOC single-platform testers such
as LTX's Fusion.
Instead of a useful tester life of five to 10 years or more, which they have
been getting with their multiple-platform testers, says Blethen, "if they
order today, they may have to switch in three years. Tester costs aren't
recovered until after the fifth year." So, he suggests, they could face
writedowns on their equipment.
Blethen's opinions, while shared by some semiconductor equipment
companies, are challenged by Teradyne and a semiconductor equipment
analyst. "Customers are voting with their orders," says Teradyne's
Newman, "and it's not even a contest. The winner in the SOC market is our
Catalyst System. [LTX's] differentiation is that they've got only one
offering and everyone else has more than one. We give our customers
whatever they want, multiple choices, not one that we think they should
want. They've got one horse. Unless they can convince the world it's the
right horse, they are going to be in some jeopardy."
Semiconductor equipment analyst Min Pang at S.G. Cowen Securities Corp.
in San Francisco agrees with Teradyne. "Anybody can give you a very
qualitative argument about why his approach is the better one, but when it
comes to quantitative arguments, that's another deal altogether.
Unfortunately for LTX, Wall Street looks at numbers and the numbers now
clearly favor Teradyne's approach."
Pang says that while there's a market for SOC, it's
very application specific. For example, chipsets
and wired handsets are a combination of silicon-
and gallium arsenide-based chips. With
telecommunications, you've got to have high-speed
data transfer. Silicon doesn't transfer data at
high speeds. The manufacturing technology of each
of these is different. So you can't get SOC in that
area. You can have sub-systems on a chip, but not
full SOC.
As a result, says Pang, it's only going to be a
modest segment of the market, and Blethen has bet
the farm on SOC. Because they've spent as much
money as they have on a single platform and SOC,
they probably will become a leader. "But the
question for LTX is how big is the market going to
be and how soon will it grow rapidly?" says Pang. He doubts it will get to
be much more than 8% of the total semiconductor test equipment market.
Understandably, LTX's Blethen appears convinced single-platform Fusion is
his company's big opportunity, just as he was filled with conviction as a
24-year-old, when he, Graham Miller and five others left Teradyne to form
LTX and develop the first mixed-signal tester. At the time, Teradyne
thought there were better opportunities elsewhere. It only caught up and
passed LTX in the mixed-signal business by the early 1990s.
Blethen is hoping that lightning will strike twice in the same place. He says:
"I believe Teradyne's management is in denial about the shift to SOC that
is starting." And this time, he says, LTX is in much stronger shape with its
technology than it was as a start-up. Moreover, as a result of the
consolidation and the sale or spin-off of some marginal operations, Blethen
says break-even costs have fallen from around $56 million in July 1998 to
around $35 million in January 1999. According to LTX CFO David Tacelli,
by July 1999, break-even should be down to $30 million, a major cost
reduction. Spurred by Fusion product gains, new orders were up 20% in the
January 1999 quarter and further sales are expected. So profit prospects
are brightening.
Allies in Asia
Moreover, Blethen says LTX now has global opportunities. He has entered
into joint ventures with three large Asian companies. The most significant
one was last April with Tokyo-based Ando Corp., a $500-million test
company 51% owned by NEC Corp., also based in Tokyo, which gave LTX
$20 million in return for the exclusive rights to sell Fusion testers in
Japan.
"Ando is developing modules of hardware and software," says Blethen, "and
they've taken all of their non-memory semiconductor test operations and
put them on Fusion. So, for the first time in the industry, two companies are
co-developing the same tester, software and hardware, and software
modules. Ando fully owns Japan. We own the rest of the world."
Giving up Japan was a high price to pay, but having access to Ando's
technical skills and a commitment to Fusion could be worth it in the long
run.
Now it remains to be seen if the long-predicted, rapid growth of the SOC
market really takes off this year. That's of considerable interest to
institutional investors such as Princeton Services, Greenway Partners L.P.,
State of Wisconsin Investment Board and Thomson Horstman & Bryant, who
own about 28% of LTX stock.
Is their patience in waiting for a turnaround at LTX finally about to be
rewarded? |
