SAN FRANCISCO -- Later this week, graphics chip maker
is expected to jump into the world of publicly traded stocks. But be careful: The graphics business can be as nasty as a game of
The six-year-old Santa Clara company is expected to debut as early as Friday, raising between $31.5 million and $38.5 million by selling 3.5 million shares at an offering price estimated between $9 and $11 a share. (On Wednesday that range was boosted from the previous $7 to $9 a share target.) The offering is being underwritten by
Morgan Stanley Dean Witter
Hambrecht & Quist
For all the allure of Nvidia as an IPO candidate, many who follow the graphics-chip industry can't help but think of the risk involved with many of Nvidia's publicly traded peers. "There's a lot of big money to be made and a lot of big money to be lost," says
C.E. Unterberg Towbin
chip analyst Claude Hazan. (Unterberg has no underwriting relationship with Nvidia.)
Nvidia plays in the same field as
. The market in general looks healthy --
expects the graphics-chip market to grow from just 4.9 million chips in 1997 to 138 million in 2001. But history has shown these are volatile stocks, having burned many an investor.
"It looks really cool," says Kevin Landis, a fund manager at
whose $52 million
Technology Leaders Fund
returned 78% in 1998. Landis has made a small killing this year with early plays on semiconductor stocks. "But look at the one variable consistent with the industry: There have always been plenty of competitors."
Landis learned years ago to stay away from a stock like Invidia. "Way back I owned
and for a brief terrifying period, I owned S3,"' he says. "I swore off them for the same reason I swore off disk drives. It's an important technology but the competition is too fierce."
"There is a lot of big money to be made and a lot of big money to be lost," says C.E. Unterberg Towbin chip analyst Claude Hazan.
If you don't believe Landis, read the prospectus. The first risk factor named is giant
, which released a competing chip, the i740, in March of last year and which has signed cross licensing deals with Nvidia's competitors,
and S3. On top of that, 3DFX recently purchased Nvidia's customer,
, a motherboard maker.
These stocks crest and plummet, says Hazan, but those who can time it right can make money. Take
, which went public in November 1996 with an offer price of 11 but rose 354% to 50 1/4 10 months later. But it has traded under 8 since May, despite a dramatic rise in the
Philadelphia Semiconductor Index
over the past four months. TDDDF closed today at 5 1/8, up 3/8.
Then there's 3DFX, one of Nvidia's top competitors. In June of 1997, 3DFX went public with an offer price of 11 and an opening price of 13 1/4. A month later it lost a key contract with
and the stock dropped to 8. It rose back to a high of 32 1/8 by April but has since declined, closing today at 13 3/8.
Graphics chips go through six-month design cycles, the quickest in the semiconductor business. From one cycle to another, one company can push another off the leader board and that can cause wild fluctuations in revenues and profits. "People are skeptical of graphics deals," Hazan says. "They will always have great blowout quarters followed by topsy-turvy quarters," he says. Nvidia has argued that it addresses that problem by having three design teams working on new products separately and simultaneously. But analysts such as Hazan say they're skeptical whether that system can smooth out cyclical design problems.
, S3 led the graphics-chip market in 1997, followed by
. But preliminary data for 1998 shows that ATI has taken the lead, and 3DFX and Nvidia -- companies previously housed further down the list -- have been gaining market share, says Dataquest analyst Georg Iwanyc. So be warned. Play Nvidia right and the stock might gain point after point. But play it wrong and it could get blasted out of the sky.