Cypress vs. IDT: They've taken different tacks to cope with their SRAM problems. Which one will prevail?
By Robert Ristelhueber
photographs by Robert Holmgren
They're often viewed as twin companies. Both were founded
in Silicon Valley in the early '80s, have pursued similar
product strategies and have wound up with almost identical
revenue. But a closer look at Cypress Semiconductor Corp.
and Integrated Device Technology Inc. (IDT) reveals
some glaring differences--including their prospects for
success.
Cypress has long attracted attention out of all proportion to
its size and importance. With flamboyant T.J. Rodgers as
its CEO, Cypress has always managed to deliver
entertainment, if not shareholder, value.
Dubbed Silicon Valley's bad boy, Rodgers has repeatedly
made headlines by slamming cherished industry icons such as
the Sematech consortium and blasting competitors that
accepted government largesse.
Yes, that was T.J. taking a network television reporter on a
tour of his wine cellar during a recent program focusing on
greed (Rodgers defends it), and T.J. dressed as Uncle Sam
holding sausages in a magazine article dealing with
government pork (Rodgers opposes it). For all his criticism
of Washington, though, Rodgers revels in hobnobbing with
the political elite. He frequently jets east to testify before
congressional committees, and has played host to at least
three senators in the past two years at Cypress' San Jose
headquarters.
But those hijinks can't conceal Cypress' mediocre
performance lately. Despite his reputation as a doggedly
hands-on manager, some have suggested that his
extracurricular activities have distracted Rodgers from
running his company. Attempts to ask him about it for this
article were unsuccessful. Rodgers cancelled a scheduled
interview appointment at the last minute to prepare for an
appearance before a congressional committee.
IDT, by contrast, has made obscurity into an art form.
Until this year the Santa Clara-based firm had shunned
publicity, and, as a result, is covered by only three sell-side
securities analysts, less than half the number writing about
Cypress. The company is sometimes confused with a
telecommunications company using the same initials.
Len Perham, the firm's CEO, resembles a kindly grandfather
rather than a Rodgers-style mogul. "We must have had
different experiences in the crib," jokes Perham.
That contrast extends to their respective management
styles. Rodgers once posted a sign above his desk that said:
"Be realistic--demand the impossible." One industry
executive with close ties to both companies calls Rodgers
"an absolute autocrat. Everything revolves around him." In
No Excuses Management (Doubleday, 1992), a book
co-authored by Rodgers, he writes: "Winning is what
matters. And if winning means being tough, demanding,
impatient, then that's what you have to be." (For insights
into Rodgers' attitude about motivation, see p.48.)
Perham, on the other
hand, is "more of a
delegator than T.J.,"
the industry
executive notes. "Len
runs his business like
a family unit. He's
interested in
camaraderie, and
sometimes protects
the weak ones in the
family."
Perham agrees that
he'd rather let
subordinates hammer
away at a problem
than impose a solution
himself. But he
concedes that, "I
have very little
patience," and has
shown that he can be
decisive as well.
In late 1994, Perham
ran into Glenn Henry,
a former IBM fellow
who told the IDT executive he could produce a chip that
would give IDT a beachhead in the desktop microprocessor
market. Though it would mark a radical departure from the
company's traditional business model, Perham seized the
opportunity and hired Henry. Then, without getting approval
from his board of directors, Perham committed $15 million
to Henry's idea. "I told him I'd find the money," he recalls.
Mutual disadmiration society
There's no love lost between Perham and Rodgers. Perham
says that Rodgers is "modeling himself as a slightly shorter
Jerry Sanders," the bombastic CEO of Sunnyvale-based
Advanced Micro Devices Inc. (AMD), for whom Rodgers
once worked. And Rodgers likely had IDT in mind when he
wrote disparagingly of "the ultimate hypocrisy of
warm-and-fuzzy cultures that don't deliver."
Cypress and IDT have one big thing in common: both have
struggled as the mainstay of their product
portfolios--static random access memories (SRAMs)--has
suffered a price collapse. A glut of SRAMs has pummelled
the sales and profits and cratered the stock price of both
companies over the past two years.
Their responses to this setback,
however, are sharply different, and
will likely determine whether one
company will be left behind.
Whereas Cypress is determined to
tough it out in SRAMs, IDT is
rapidly backpedaling away from the
market. As a result, the perception
of the two firms as virtually
identical twins may soon be history.
In Cypress' 1997 annual report,
Rodgers poses the question many
shareholders have been asking: "Why not just get out of the
SRAM business and focus on your most profitable business
segments?" He responds that although such a move might be
advantageous for a single year, it would be a long-term
mistake. "[When] the SRAM business makes money, it makes
a lot of money," he writes, noting that SRAMs have
contributed roughly 60% of the company's profits over the
years.
"Our efforts to diversify the company into other product
areas have actually been funded by profits from the SRAM
group," he claims. Thus, Rodgers issues his marching orders
to the troops: "Making the SRAM business profitable again
is our primary challenge."
But Perham views that as a fool's errand. "It became clear
to us back in 1995-96 that because Intel was hyping the
fact that the world was going to need an infinite supply of
SRAMs, everybody who could drive a nail would build an
SRAM plant, and there was going to be tremendous capacity
available.
Are SRAMs commodities?
"It looks to me
like the SRAM
business is going to
be under duress
for the next 10
years," Perham
asserts, challenging
Rodgers' belief
that profitability
is right around the
corner. IDT has
already nearly
abandoned the
price-sensitive
cache SRAM
segment, and is
focusing on the
niche aimed at
communications
segments.
Perham projects
that SRAMs, which
represented about
32% of the
company's sales in
the last quarter,
will fall to just
25% of revenue
during this fiscal
year (it was nearly
50% a couple of years ago). In contrast, Cypress derived
43% of its revenue from SRAMs in the first quarter of
1998.
Given his druthers, Perham would probably jettison SRAMs.
"Memories are always treated as a commodity by investors.
You can't provide exciting returns for your shareholders
when much of your business is competed for on cost."
But many customers would oppose that move, including
important communications equipment vendors like Cisco
Systems and Bay Networks who buy other types of
components from IDT. IDT also wants to retain its memory
expertise because microprocessors, one of its key products,
contain large memory elements. Perham might shift SRAM
manufacturing to outside foundries to bolster profits,
though.
Cypress is taking a batten-down-the-hatches approach to its
SRAM woes. In March, the company closed one wafer fab
and downsized another, consolidating production at a more
modern facility in Minnesota. The move, which resulted in
the elimination of 100 jobs and a charge of $51 million, was
criticized by at least one analyst as long overdue. Rodgers
said the changes "will significantly reduce our costs and
improve our efficiency." He also vowed to introduce more
value-added SRAMs to the portfolio to fatten margins.
The author of No Excuses Management offers this excuse
for last year's performance: "[We] were disappointed that
the SRAM slump continued unexpectedly for another full
year."
Some speculate that, unlike IDT, Cypress has no choice but
to slug it out in SRAMs. "IDT is doing a pretty interesting
job in trying to get into higher-margin logic business," notes
Mark Edelstone, financial analyst with Morgan Stanley in
San Francisco. "Cypress is trying to reduce their
dependency on SRAMs as well, but it's been a bit of a
struggle."
Cypress' recent history is replete with diversification
efforts that have misfired. Its microprocessor subsidiary,
Ross Technologies, was sold in 1993 after relations soured
with Sun Microsystems Inc., the company that had licensed
Cypress to make the chips. A two-year attempt to enter the
chipset business was scrapped because Cypress failed to
execute quickly enough. And an ongoing effort to compete in
the high-density programmable logic market has made scant
headway thus far. Rodgers concedes in the latest annual
report that the division "has not grown as rapidly as we
would have liked."
The company has been handicapped because its process
technology is generally not leading edge. This was glaringly
obvious in its struggle to perfect a six-transistor cell
technology. Because that project was delayed by more than
a year, Cypress has often been stuck with die sizes that are
larger than competitors' parts.
Cypress' diversification scorecard isn't totally bleak,
however. The company has made headway with data
communications products, which represent its
fastest-growing group. In computer products, the company
has high hopes for its universal serial bus (USB) chips,
particularly now that Microsoft Corp., Redmond, WA, is
about to release its Windows 98 operating system, which
will support USB.
Drew Peck, a security analyst with Cowen & Co., Boston,
recently upgraded his recommendation of Cypress' stock
from "neutral" to "buy" partly because of the potential of
USB. "I think Wall Street has underestimated the USB
phenomenon, and Cypress will be a major player there," he
says. But Cypress may only benefit from fat USB margins
for a year or so, Peck warns: "Ultimately, that will become a
pretty grim commodity product."
Winchip windfall
IDT has been less dependent on SRAMs in part because it
made the correct bet on microprocessors a decade ago. By
licensing the MIPS architecture, the company has
benefitted from the success of that design in embedded
control markets.
But most of the buzz about IDT lately stems from its
plunge into the desktop microprocessor market in
competition with Intel, AMD and National Semiconductor.
The company's Centaur Technology subsidiary, the one
headed by Henry, last fall introduced the WinChip, a part
aimed at the sub-$1,000 desktop personal computer market.
Perham says IDT shipped
between 100,000 and 200,000
units of the WinChip in the
first quarter of this year, and
expects WinChip revenue to
represent over 10% of company revenue during this quarter.
The part is small and easy to make, he says. When IDT
starts building it using a 0.25-micron process, the die will
be cheaper than the package holding it, Perham claims.
For the moment, IDT is steering clear of Intel's wake. The
firm is concentrating on third-tier manufacturers using
distributors, and so will primarily compete with National's
Cyrix unit. Its best-known customer so far is Trigem
America Corp., a Korean-based PC maker that sells in the
United States through Costco stores.
But Perham is looking to higher plateaus. He says he has
sampled the chip with many first-tier vendors, and the
company recently signed Armonk, NY-based IBM Corp. as a
foundry to satisfy larger customers who want the assurance
of an alternative to IDT's fabs.
IDT has a decent shot to succeed with WinChip, contends
Dean McCarron, an analyst with Mercury Research in
Phoenix. Given the $20 billion size of the market for
desktop processors, he says, IDT needs to capture only a
tiny sliver to get a big boost from the WinChip. Investors
share that optimism. IDT's stock has risen sharply this year
in anticipation of WinChip's success.
Although the companies are today nearly equal in size,
IDT's prospects look brighter because its diversification
efforts have panned out a lot better than those of Cypress.
Perham boasts that IDT's strategy to pursue higher-value
products will leave Cypress in the dust.
Rodgers contends that, "Most companies don't fail for lack
of talent or strategic vision. They fail for lack of
execution." A year or so from now, the market will show
whether it was Cypress or IDT that put the right strategy
and execution together in time. |
