Transmeta: Are the chips down?

The chip company, with its pioneering designs and concepts and high-profile customers, was once charged up to take on Intel. But things haven't turned out quite as planned.

Michael Kanellos Staff Writer, CNET News.com
Michael Kanellos is editor at large at CNET News.com, where he covers hardware, research and development, start-ups and the tech industry overseas.
Michael Kanellos
9 min read
Toshiba had big plans earlier this year for the U.S. release of a laptop with an energy-efficient chip from Transmeta, a product that would have represented a coup for the upstart chipmaker in its rivalry with mighty Intel.

But in the latest setback for Transmeta, Toshiba quietly pulled the plug on the notebook this summer, partly because of delays surrounding Transmeta's latest chip, executives told CNET News.com.

Since it launched a successful public offering one year ago, the company has switched CEOs twice, the stock price has plunged amid the broader economic slump, and its chips are still scarce in the United States. In addition, the company has been hit with manufacturing missteps that have prevented it from stealing customers from Intel and carving out a successful niche in the $22 billion microprocessor market.

In some respects, Transmeta's battle with Intel mirrors the plight of smaller companies that have competed with Microsoft in the software market. In both industries, the barriers to entry have become so high that many companies are reluctant to compete, and those that do fight a David-and-Goliath battle that has been brutal on the little guy.

"Transmeta is not relevant," Jerry Sanders, CEO of rival Advanced Micro Devices, asserted in a recent conference call. "Intel has effectively kept them out of the marketplace."

A Transmeta representative defended the company's Crusoe chip and its track record, which includes early deals with high-profile customers such as Sony, Casio, Fujitsu and NEC. The company "continues to aggressively pursue major U.S.-based customers and believes that these efforts will result in broader adoption of Crusoe in the United States," the spokesman said.

Nevertheless, its global market share remains almost imperceptible. The stock trades at under $2 per share, down about 96 percent from its all-time high of $50.88, which it reached on the day of its initial public offering in November 2000.

Concept vs. execution It wasn't supposed to turn out this way. Proponents of Transmeta, which was founded in 1995, believed the company had learned from the failures of other would-be Intel rivals, from Cyrix to Centaur. Rather than compete with Intel on price, Transmeta intended to create a superior product that would quickly capture a small, but profitable, segment of the market--much in the way Apple Computer has survived with less than 5 percent of the operating system market.

Transmeta falloff
Stock price from November 2000 to present.  

Source: Prophet Finance
Specifically, Transmeta executives planned to create energy-efficient chips for laptops and other computing devices.

Unlike Intel's processors, Transmeta's Crusoe chips weren't power-hungry behemoths that would last only a few hours in a battery-powered laptop. Crusoe promised to use less power through a novel architecture that would shift processing functions from transistors to "code morphing" software.

The code-morphing technique translates computer instructions meant for an Intel x86 chip into ones the Transmeta chip understands. Once translated, the instructions live in a special patch of memory that is easily accessed. Improvements to the translation software theoretically improve the performance of the chip.

For example, a 700MHz Crusoe could run Windows programs as fast as a 500MHz Pentium III--the best Intel offered in January 2000. In addition, Crusoe chips were smaller and cost half as much as their Intel counterparts.

Unfortunately for Transmeta's backers, the concept has proved easier said than done.

The chips, announced at Transmeta's glitzy coming out party in January 2000, didn't appear in notebooks for another 10 months. In fact, manufacturers never adopted the Crusoe 5400, the company's first chip, but instead went straight to its successor, the 5600.

The first Crusoe notebooks most U.S. analysts saw were models that Transmeta founder and then-CEO David Ditzel brought to the Microprocessor Forum in October 2000. Ditzel passed out notebooks for quick inspection in the hallway, but analysts didn't get evaluation units for more thorough testing for another month--until after the company's IPO.

After the IPO, NEC recalled nearly 300 notebooks containing Transmeta's chip because of a manufacturing flaw. Though it was promptly fixed, the defect made it difficult to reinstall operating systems. Days later, Sony warned Japanese consumers that some of its Vaio notebooks might have problems because of Transmeta chips.

Meanwhile, Transmeta was courting Taiwan Semiconductor Manufacturing Co. to produce its next chip, the Crusoe 5800. IBM had been making the chips, but Transmeta wanted a lower-cost manufacturer. In February, Transmeta struck an exclusive deal with TSMC.

But the switch didn't end the delays. Samples of the 5800 chip that Toshiba received had problems, which seemed destined to push the project to November and prompted Toshiba to kill the notebook for the U.S. market.

"We'd get products and then find an anomaly. You can put in a workaround but the only way to fix it is through silicon," said Steve Andler, Toshiba's vice president of marketing.

Intel outfoxed them. Before he was forced out last month, Transmeta CEO Mark Allen said the company was still completing "long-term operating life" tests on the 5800.

Sources familiar with the situation said that some of the problems stemmed from the complex design of the chip as well as from Transmeta's testing procedures, which were not weeding out inadequate chips but were giving the company an early, erroneous impression of success. Others, however, blamed TSMC's manufacturing processes. Early on, many of the faulty chips consistently came from the same section of the wafer, which sources said indicated a manufacturing flaw.

Normally tight-lipped TSMC blames the 5800's design.

"Outfoxed" by Intel
To some extent, the blame game is pointless. Heated competition from cross-town rival Intel simply means that Transmeta has little margin for error.

Intel has cutting-edge factories that produce chips more efficiently than anyone else's. With 12 wafer fabrication sites and 12 separate assembly and testing plants worldwide, Intel has proven adept at wiping out rivals: Out of a half-dozen that have tried, only Advanced Micro Devices of Sunnyvale, Calif., survives independently--and AMD still routinely loses money. Others, including Rise Technology, Cyrix, National Semiconductor, IDT and Centaur, have gone out of business or sold out to one other.

To counter Transmeta, Intel cranked up the engineering and marketing pressure.

Between January 2000--when Transmeta introduced the chip--and the delivery of Transmeta's first chips nine months later, Intel started work on a new line of energy-efficient processors and began to promote them heavily with major computer makers such as IBM.

Several PC makers, including IBM, Compaq Computer and Dell Computer, purchased Intel's low-power chips rather than Transmeta's. As recently as Tuesday, Sharp said it would stick with Intel's energy-efficient chips for its notebooks sold in the United States, although it sells a Transmeta notebook in Japan.

On the marketing side, Intel changed the way it talked about its chips. Instead of stating the "worst case" power consumption of its processors--a measurement that engineers use to discern how much power a processor needs at full speed--the company began emphasizing average consumption.

As a result, Intel was able to claim its chips required 2 watts of power, compared with 1 watt for Transmeta, said Nathan Brookwood, principal analyst at Saratoga, Calif.-based Insight 64. Most analysts considered the difference marginal.

"That took the wind out of Transmeta's sales," Brookwood said.

As a final blow, Intel lowered prices on its notebook processors to ensure that Transmeta would be relegated to mini-notebooks, a sliver of the overall notebook market.

"Intel outfoxed them with the full-factor stuff," said Joe Osha, an analyst at Merrill Lynch. The Crusoe "is a neat, small-form-factor product, but it is not an Intel killer."

Overseas troubles
As Intel hammered away at Transmeta, pressure also mounted across the Pacific.

Transmeta's biggest customers--including Toshiba, Fujitsu and NEC--are in Japan, where consumers have long favored smaller, more energy-efficient notebooks. But a sustained recession has dented demand in Japan, where unemployment is at a record 5 percent. Some company executives attribute the steep drop in Transmeta's revenue--from $18.6 million in the first quarter to $5 million in the third quarter--to the Japanese economy.

It is a slice of a slice of a slice of a pie. In April, Transmeta was on track to ship 1 million chips in 2001. But so far this year, it has sold about 500,000 chips, in large part because of the Japanese demand slump, said Dean McCarron, principal analyst at Mercury Research. Companies need to hit about 1 million units in annual volume just to stay in the market, he noted.

In addition, some new products that were using the Crusoe have been canceled. Gateway and America Online adopted a Crusoe chip for a new series of Internet appliances last November, but lackluster sales led to the demise of the product line.

The U.S. economic slump also has hurt Transmeta. FiberCycle and Rebel.com, two start-ups that hoped to popularize skinny, energy-efficient servers with Crusoe processors, gave up on Transmeta-based products this year.

The deteriorating economy and Intel's successes have created turmoil in Transmeta's executive offices. The company has gone through two chief executive changes this year.

In March, Ditzel stepped down as CEO to become vice chairman and chief technology officer. Ditzel, who sold 400,000 shares of Transmeta stock in May with a market value of $4.9 million, previously directed Sun Microsystems' Sparc Laboratories and engineered chips at AT&T's Bell Labs.

Allen, then president and chief operating officer, replaced Ditzel as CEO. A veteran of Nvidia, C-Cube Microsystems and other companies, Allen was supposed to bring greater efficiency to Transmeta.

But many rank-and-file workers complained that Allen was too rigid. And two of the seven founders--Ed Kelly and Colin Hunter--left during Allen's reign.

In mid-October, the board shook up its management team again. Chairman Murray Goldman became interim CEO, and board member Hugh Barnes filled the roles of president and chief operating officer.

Out of sight, out of mind
Although sources said that Allen's departure has improved the mood at Transmeta, the executive shake-ups have caused analysts to question the board's leadership. That has eroded what was once Transmeta's chief advantage: buzz. The company is falling off the industry's radar screen.

IDC estimates that the company's chips accounted for 0.3 percent of processors produced from the fourth quarter of 2000 to the present, or about 2 percent of the market for notebook processors .

"They've got a part of the market that is interesting. The question is whether they can generate the volumes that will make it all worthwhile," said Kevin Krewell, an analyst at the Microprocessor Report. "It is a slice of a slice of a slice of a pie."

It is difficult to predict what will happen to Transmeta if its stock price continues to decline and manufacturing delays persist. Five years of operating in ultra-secret stealth mode has made Transmeta veterans relatively media shy; few executives have articulated a clear strategy to Wall Street. Transmeta has about $262 million in cash, but it expects to burn through $20 million in the current quarter.

"It's been just over a year since Transmeta shipped its first microprocessors and the top seven Japanese notebook manufacturers have introduced Crusoe-based systems--and four of these customers have already brought (notebooks) to the U.S. market," a company representative said in an e-mail. "By any fair analysis, Transmeta has established a sound customer base in the short span of about a year shipping microprocessors, particularly in a down economy."

Despite that optimism, Transmeta's stock price makes some people think it's an obvious acquisition target. Several sources speculate that Taiwan's Via Technologies, which competes against Intel, could acquire it.

But even with a market value of about $235 million, Transmeta would be a risky purchase: The company's chief asset is its intellectual property, and keeping it would require retaining the engineering staff. That could be tough to do with a group of people who have bristled at previous corporate shake-ups.

"Right out of the blocks, I would suspect that some of the employees would walk out. Most of the smart guys have money, and they are motivated by other things," said Brian Alger, an analyst at Pacific Growth Equities. "These are quirky, quirky people."

Independent survival could be equally challenging. Although a number of new Crusoe notebooks from Sony and other Japanese PC makers are slated to come out this month in the United States, pressure from Intel and the dour economy will likely keep a lid on sales.

"The execution is pretty vital, especially if you are going to compete against Intel," Osha said. A turnaround "is conceivable, but it is going to be a major challenge."