IBM Jilts Transmeta at IPO altar
IBM Jilts Transmeta at IPO altar
By Hal Plotkin
Silicon Valley Correspondent
If you think of an initial public offering as a kind of wedding between investors and a company, it looks like much-talked about Transmeta Corp. has just been jilted at the altar.
Think of it this way: Just as the minister was about to ask whether anyone could say why these two should not be joined in holy matrimony–International Business Machines Corp., {IBM} today raced to the altar with this startling news: It’s dumping plans to use Transmeta chips to power its very popular ThinkPad laptop computers.
The news could deal a crippling blow to Transmeta’s hopes for a successful public offering, IBM’s lack of support is particularly troubling considering that IBM manufactures the chips in question for the upstart would-be rival to Intel Corp. Officially, Transmeta is still planning to take the company public on Nov. 6th.
Even before the news, Transmeta Corp. already had a good chance of walking away with this year’s award for “most hype generated by an unproven pre-initial public offering company in a tiny market niche.”
You’ve no doubt already heard a lot about Transmeta, which is easily one of Silicon Valley’s most closely watched almost-public companies. The firm made its first big splash by hiring the famous Swedish Linux software guru and namesake, Linus Torvalds, and then sealing him to well-publicized secrecy about his labors.
The gambit worked. After all, everyone loves a secret.
The showmanship finally reached a crescendo early this year when Transmeta pulled back the curtains on what it says is a revolutionary new family of microprocessors that it claims will make batteries last up to a full day longer in notebook computers and similar devices. Intel Corp. {INTC} ought to be quaking in its boots, the company’s backers suggested.
Buoyed by the attention, Transmeta took what looked like a victory lap, landed on magazine covers and front pages across the country and then began readying itself for the big IPO.
Transmeta’s filing indicates that it plans to raise up to $156 million by offering 13 million shares priced between $10 and $13 each. Transmeta has more than $218 million in cash, cash equivalents and working capital on hand, according to its pre-IPO filing with the Securities and Exchange Commission, though the company uses a figure of $500 million to include, on a pro forma basis, the proceeds of its not-yet-happened IPO.
As of June 30, Transmeta had an accumulated deficit of $119.4 million, which includes losses of $10.1 million in 1998, $41.1 million in 1999 and $43.4 million in the first half of 2000, on revenue of $358,000. The company attributes most of those losses to product development, sales and marketing and administrative expenses.
IBM’s announcement illustrates one of the main problems with the pending IPO, however, one that several leading analysts have been harping on for months.
They say that Transmeta has yet to prove that it can capture even a 1% market share over the next year, which would bring total revenue to just about what 15 busy neighborhood supermarkets do, or about $250 million. And that would happen only if everything goes just right, which today’s news indicates is not the case.
Transmeta’s prospects rest on whether its new low-power processor, called Crusoe, has all the advantages the company claims and whether the upstart firm can hold off growing competition over time in its target niche from the likes of established players such as Intel and, quite possibly, Advanced Micro Devices Inc. {AMD}.
To be fair, there are at least some indications that Transmeta has a fighting chance.
While the news from IBM is a blow, several other customers, including Fujitsu Ltd. {FJTSY} and Sony Corp. {SNE}, which is also a Transmeta investor, have agreed to purchase undisclosed quantities of the new chip after making their own comparisons with other products available from Intel.
Several other companies, including Gateway Inc. {GTW} and Hitachi Ltd. {HIT}, have also said they plan to use Transmeta’s chips in forthcoming products, though volume shipments to those customers hadn’t taken place as of the end of September. Gateway is also a Transmeta investor.
The bigger question, however, is whether sales of the magnitude generated by current order flows will lead the stock to a performance that is commensurate, over time, with the considerable enthusiasm that is expected to greet what is expected to be a very flashy IPO.
“As a conservative investor myself, I would definitely wait a little longer,” says Josephine Mong, semiconductor research analyst at market researcher International Data Corp. “Nobody has been able to get a hand on Transmeta’s chip yet so we don’t really have any independent third-party public benchmarks to back up what they’ve been saying.”
Transmeta officials say the existing industry benchmarks, which they eschew, were originally designed to assess the performance of desktop computers that are plugged into the wall and consequently don’t accurately measure the true advantages of the company’s unique variable power-consumption technology.
Unlike competing products, Transmeta’s chip, it is claimed, is smart enough to know exactly how much power it needs for each particular operation executed, which allows it to reduce the amount of power drawn from batteries and thus extend their life.
What’s more, Transmeta’s low-power chip is designed to run programs based on Microsoft Corp.’s {MSFT} Window’s operating system, a feature the techies call “x86 compatibility.” By contrast, the low-power chip Intel has positioned to attack the mobile, handheld market is built to be compatible with Microsoft’s mobile-computing platform, called Windows CE.
Transmeta’s design makes it possible for mobile-Internet users to take advantage of the entire content of the Internet as it currently exists rather than just those Web sites that have converted their content into the much less-popular Windows CE format.
“X86 capability is very important,” Mong says.
That is one of the reasons industry analysts are generally, though cautiously, impressed.
But some just as quickly add that, as far as power conservation goes, components such as liquid crystal display screens and hard drives, not just the microprocessor, are also responsible for draining the life out of batteries in mobile-computing devices.
“I’m skeptical of Transmeta’s claims,” says Nathan Brookwood, an analyst at Insight64, a Saratoga, Calif.-based market-research and consulting firm that has previously done work for both Intel and Advanced Micro Devices.
“The microprocessor accounts for some fraction of the power consumed,” Brookwood says. “Nobody knows for sure, but I’d estimate the overall savings might provide a 10% or 20% improvement in battery life, although that still remains to be demonstrated.”
Like Brookwood, several other leading industry watchers also say it is much too soon to know whether Transmeta’s assertions will be born out in the marketplace, particularly if Intel succeeds with its already announced plan to fend off its newest competitor by bringing a rapid succession of its own similar new products to market.
At present, the firms are fighting a rather heated press release war, with each side claiming competitive advantages for their own current and forthcoming mobile-computing chips.
“Transmeta has some early design wins and has a chance to be dominant in what is maybe a 5% market niche,” says Martin Reynolds, research fellow at the Gartner Group’s Dataquest. “But it’s still quite early, and you can never tell for sure.”
To be fair, Reynolds says that capturing 5%, or even 1%, of the $25 billion to $30 billion annual market for microprocessors would be nothing to sneeze at. “They don’t have to be very successful to pay their bills,” he says, noting that 1% of $25 billion is $250 million.
“One percent market penetration in the first year seems a bit high to me,” Mong counters. “They could get there, but honestly, I don’t see it happening in the next year.”
Rob Enderle, vice president of the Giga Information Group in Santa Clara, Calif., says that Transmeta has cracked open a window of opportunity in the emerging mobile-Internet-appliance market.
But he quickly adds that window could also slam shut pretty quickly if Intel, which, at least until quite recently, has had its attentions focused elsewhere, gets its lower-power act together as it has now promised.
“Intel has alienated a lot of the hardware [original equipment manufacturers] by not listening to them,” Enderle says. “Some of them are eager to do business with Transmeta because it’s not Intel and because it gives them somewhere else to go.”
Enderle says Intel’s chip designs have traditionally been centered on desktop computers and servers, which account for the vast majority of the market that now exists. Vendors of smaller, more-portable computing products have been clamoring for some time now, he says, for solutions that are more suited to their needs.
“Transmeta paid attention to that, and Intel didn’t,” Enderle says.
Like the other analysts, Enderle says he would still like to see third-party benchmark tests that verify the reputed advantages of Transmeta’s new chips. But he adds that the recent decision to buy chips from Transmeta by vendors such as Sony, which is known for its careful evaluation of components, says a lot.
“Sony was able to look at both [Intel and Transmeta] chips and judged Transmeta’s superior,” Enderle says. “That’s important because it’s extremely costly for someone to go with something other than Intel.”
The industry analyst notes that computer sellers who switch to non-Intel products lose out on the considerable benefits of the “Intel Inside” cross-marketing promotion and other support services offered by the world’s leading maker of microprocessors. PC makers, in particular, have been extremely reluctant to antagonize Intel in the past, given the demand for products containing the firm’s microprocessors.
Some analysts have speculated that IBM’s announcement may have been a result of Intel making the company a better offer.
Enderle thinks Intel could rather quickly take back the mobile market, at least some of which now appears to be slipping through its fingers.
“If Intel had been paying attention to this space, Transmeta would not have been able to get in the door in the first place,” Enderle says. “It’s possible Intel could mend those relationships, though, and if that happens it would cause Transmeta a lot of grief.”
Martin Reynolds, of Dataquest, thinks it is unlikely that Intel will do what it needs to do to keep Transmeta from achieving at least some small measure of success.
“Intel has a response and will have other responses,” Reynolds says. “But as long as Transmeta stays focused it probably wouldn’t make sense for Intel to try to be dominant in that one small part of the market.”
In essence, Reynolds says, it just isn’t worth it. Not if Intel risks taking its eye off more-important and more-lucrative markets by fighting a small stakes war in a head-to-head battle with Transmeta. But he also says it would be a mistake to permanently count Intel out of any fast-growing microprocessor market segment, given the firm’s pattern of fierce and tenacious competition.
“Transmeta definitely has the hype going for them; they’ve done a great job on that,” Mong adds. “But I find it very hard to believe that Intel would neglect what will become a very important part of microprocessor designs in the future.”
The firm’s pre-IPO investors include America Online Inc. {AOL}, Compaq Computer Corp. {CPQ}, Deutsche Bank AG {DTBKY}, Five Points Capital, Gateway Inc. {GTW}, Invemed Associates, Phoenix Technologies Ltd. {PTEC}, Samsung, Sony, Soros Fund Management, Tudor Investment Corp., Van Wagoner Capital Management, Paul Allen’s Vulcan Ventures.
