3Dfx Bull's Pen
Louis Corrigan (TMF Seymor)

Since my foe Jeff Fischer helped write the persuasive 3Dfx Fool Portfolio buy report, I feel like a bull in a china shop: a bit clumsy by comparison but happy to crash the party! The bull case comes down to three points. First, 3Dfx has got the best technology in one of the only areas of PC-related hardware where this really matters. Second, 3Dfx has used its technology edge to cultivate the top brand in the 3D graphics chip market. Finally, the combination of great technology and a hot brand have positioned it to become a serious contender in the mainstream 3D market, which is exploding. The growth ahead should be breathtaking.

Performance rules the computer game market largely because it's driven by active gamers who pick up the fastest 3D chips available in the form of graphics add-on cards. And 3Dfx's new Voodoo 2 is the performance king. In a recent review of accelerator chip options, John Carmack, the genius behind the wildly popular game Quake, said that the Voodoo 2's numbers "were far and away the best ever recorded, and they are going to get significantly better." Voodoo 2 should "remain the highest performer for all of '98," he added.

Performance explains why gamemakers looking to create ever more intense products love the company's technology and have written more than 150 games for it. More than half advertise the 3Dfx logo on the packaging for their games. Some retailers and gaming magazines now devote the kind of attention to 3Dfx that they do to Nintendo 64, Sega Saturn, and Sony PlayStation. Such enthusiasm has built the brand while suggesting that 3Dfx can maintain its dominant position in the gamer market.

Yet that highly profitable niche is rubbing up against the blossoming mainstream market for solutions that integrate 2D and 3D graphics functions on one chip. Witness Intel's <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: INTC)") else Response.Write("(Nasdaq: INTC)") end if %> recent introduction of the i740 integrated chip. Intel may be the bogeyman of all other chipmakers, and reports from the likes of Mercury Research suggest that the i740 is strong enough to muscle out a 20% to 25% market share in the mainstream PC market. But while currently the fastest integrated chip, the i740 is slower than 3Dfx's Voodoo 2 and thus should be slower than the Banshee, the company's integrated 2D/3D chip expected in the June quarter just in time for the holiday season.

With capacity constraints easing and Asia's troubles seemingly improving, 3Dfx's chances of finding new manufacturing partners, if needed, have improved. The company's exceptional growth should continue for at least the next two years. Fourth quarter results showed sales up 123% from the third quarter, to $22.3 million. That pushed FY97 revenues to $44.1 million. Net income for the quarter rose to $2.1 million, or $0.10 a share on a fully taxed basis, crushing estimates. One might say "Yowza!" at seeing the stock (at $24) trade at 60 times the earnings per share run-rate, but UBS Securities analyst Charles Boucher projects sales will soar to $149 million this year, good for fully taxed EPS of $1.01, or double the consensus earnings estimates floating around in December. Sales should jump to $188 million in FY99, leading to EPS of $1.33.

A few things are clear from this. First, the analysts have been too conservative. Second, the company should become much more profitable, with net margins rising to 10.3% this year and 11.1% in FY99 from just 6% last quarter thanks to the relatively fixed costs of the company's fabless operations. Third, 3Dfx now trades at about 24 times 1998 earnings estimates, which assume 153% growth over even its run-rate, giving us a modified PEG of 0.39. Fourth, based on one analyst's long-term growth estimate of 30%, the stock has a YPEG fair value of $30 with a two-year target of about $40, suggesting 25% to 29% annualized price appreciation from here. Given number one and number two, these projections could prove conservative.

Of course, this leaves out the dilution from the recently announced secondary offering of at least two million new shares. That will immediately push the share count just above 16 million and reduce the above EPS projections to $0.89 for FY98 and $1.18 for FY99. Yet even after this dilution, 3Dfx trades at just 2.8 times projected '98 sales, hardly pricey if profit margins end up in double-digits. Plus, assuming the offering goes off at $24, it could swell the company's coffers to $83 million, or over $5 a share in cash versus no long-term debt. Also, methinks that swelling cash is needed to ramp up a whole lot of chips, probably more than the current estimates account for. I'd go on, but it doesn't seem sporting to use any more of Jeff's own arguments against him.

Next: The Bear Argument