Daniel, there maybe some confusion with 3dlabs and TDFX: the same thing happened last year in the summer at about this time I believe when 3dlabs preannounced some bad news.TDFX went down with it and was back up 2 days later.
Here is the recent WSJ news on S3. ( I don't know if it's been already posted ):
TA
===========================================
July 2, 1998
A Bitter Feud With Rival Intel
And Some Bad Moves Batter S3
By DEAN TAKAHASHI
Staff Reporter of THE WALL STREET JOURNAL
SANTA CLARA, Calif. -- When the U.S. Federal Trade Commission began investigating Intel Corp.'s business practices last fall, executives at a scrappy Silicon Valley chip maker called S3 Inc. lined up at the door.
Intel, whose microprocessors define many technical standards in the personal-computer industry, had refused to license a key technology to S3 and other makers of PC graphics chips. As a result, S3 scrapped plans for an important product called Plato, which had cost nearly $10 million to develop and seemed likely to bring in $100 million in revenue. With Intel becoming a major supplier of graphics chips, S3 feared that the chip giant could increasingly use its market muscle to knock the smaller competitor out.
So S3 complained to the FTC. "We figured things couldn't get any worse," says Ronald Yara, co-founder and a senior vice president at S3. "We just didn't want to sit around and wait."
Months later, relations have soured between S3 and Intel, and S3 is indeed in sad financial shape. Intel executives deny competing unfairly in graphics chips, insisting that they are under no obligation to share proprietary technologies without getting adequate compensation.
A Few Wrong Turns
But even Mr. Yara concedes that many of S3's wounds were self-inflicted, as the company fell behind rivals in a race to bring more dazzling images to PC screens. The company, which once dominated its field, has lost some $40 million on operations in the past two quarters. Its experience exemplifies how a few wrong turns can prove deadly in fast-moving markets, and the difficulties of staying focused when an industry leader starts making waves.
The S3 story also suggests challenges for the FTC. The agency has already sued Intel for the way it dispensed information to three computer makers, but is looking for other evidence of questionable behavior as Intel has moved into adjacent markets such as graphics. But events surrounding S3 seem ambiguous.
For years, graphics chips were largely ignored by Intel and the government alike. Mr. Yara, a Hawaiian whose grandparents emigrated from Okinawa, hooked up with veteran Filipino-American engineer Diosado Banatao to start S3 in 1989. They exploited several changes in the market, introducing a chip called Virge that integrated rudimentary three-dimensional graphics and fit into the same slot in a circuit card as earlier 2-D chips. Virge propelled 1996 revenue to $439.2 million and put S3 in the No. 1 spot in graphics.
The same year, S3 took what seemed like an imperceptible step off the innovation treadmill. Terry Holdt, then S3's chief executive officer, authorized engineers to create a slightly different version of Virge using a different kind of memory circuitry to store data for the chip. It turned out to be unpopular, and the project stalled.
Resources Diverted
Mr. Holdt's decision diverted resources from what should have been S3's main effort: designing a next-generation 3-D chip that would have replaced its top-selling chip. Gary Johnson, Mr. Holdt's successor, also diverted engineering resources into ancillary projects in the name of product diversification.
One key effort was Plato, an innovative attempt to save money for PC makers by combining a graphics chip with a "chip set" that controlled other data traffic within the PC. At least five customers were interested, S3 says, but some wanted a letter from Intel expressing its opinion on the technology.
Separately, Intel was in the process of pushing the industry toward using a new kind of data pathway that connects Intel's Pentium II microprocessor to graphics and the rest of the system. Information about that "P6 bus" would be critical in any successors to Plato, and Intel wasn't inclined to license it.
In August, 1996, Mr. Johnson visited Intel's headquarters and pleaded for a license, to no avail. Intel spokesman Chuck Mulloy says Intel took a uniform approach to all comers and decided not to give away the technology for free; rather, it would license it in a "value for value" trade for other technology. S3 apparently had little to exchange.
Mr. Mulloy says Intel expressed some technical concerns with S3's approach, but otherwise did nothing to discourage S3 or anyone from trying solutions such as Plato. In any case, potential customers of Plato either got spooked or tired of waiting for the product to be completed.
S3 tried other less-than-successful diversions, including chips for sound cards, video and portable computers. It also spent money on a flashy new logo and mirrored-glass headquarters.
Meanwhile, ATI Technologies Inc. and Nvidia Inc. pushed ahead into new generations of 3-D chips. When S3 finally assigned several teams to follow suit in late 1996, the rivals had a big head start.
Intel further roiled the market with what it called the "accelerated graphics port," a more efficient way of moving graphics data through the PC. S3 scrambled to set up three chip-development efforts to work with the technology. One project finished on time, but two completely missed their deadlines. The more the company fell behind, the more it had to modify its designs to match rivals" chips.
Slow Motion 'Train Wreck'
"We realized our mistake, but it was like watching a train wreck in slow motion," said Neal Margulis, former vice president of research and development at the Santa Clara, Calif., company. "The competitors started earlier."
The company hit the wall in the fourth quarter of 1997. The company reported a loss of $8.1 million and was forced to restate revenue downward by $58 million for several previous quarters due to errors recording revenue at international distributors. Mr. Johnson resigned and took a post as vice chairman. Mr. Holdt returned from semiretirement to take back the top job.
"I felt some responsibility for the state of the company," said Mr. Holdt, who has a deep, stern voice. "There is a legacy a president leaves behind. Mine wasn't as strong as possible."
Mr. Holdt has laid off workers, sold some assets and repriced options to stop engineers from defecting. He acquired graphics patents that allowed him to sue one rival, Nvidia, and microprocessor patents in hopes of bringing Intel to the technology-licensing table, with no luck so far.
One of S3's key chip projects, called Savage3D, was finally completed and has received kudos in the trade press for its high performance. The company also promises new chips for mobile and desktop computers in the second half.
"It's going to be a hard road back to profits for S3," said Elias Moosa, an analyst at BankAmerica Robertson Stephens in San Francisco. "They have to offset a lot of declining revenue in their commodity parts at a time when pricing pressure is intense."
Mr. Holdt isn't predicting profits until next year. Rivals have picked off S3 engineers, and Intel, which to a large degree still controls much of the future in graphics technology, has flooded the low end of the 3-D chip market with its own products. The lesson is clear for high-tech companies sitting atop an industry.
"If you miss one generation, you're in trouble," said Mr. Margulis. "If you miss two, you're dead."
|
