Transmeta's future may not be in the chips

Transmeta's future may not be in the chips

Later this month, Transmeta will announce the results of an internal evaluation of its business model that could signal the company's gradual exit from microprocessor sales.

Transmeta warned investors last August in a filing with the U.S. Securities and Exchange Commission (SEC) that if it continued to lose money, the company would have to consider licensing its low-power technologies as its primary business activity. Last week the company said it now definitely plans to increase its licensing activities, and that it would "complete a critical evaluation of the economics of its current business model of designing, developing and selling x86-compatible microprocessor products."

The company has already signed licensing deals with NEC Electronics and Fujitsu for its LongRun2 power management technology. LongRun2 allows a processor to adjust both clock speed and voltage thousands of times a second in order to control power consumption depending on the processor's workload.

In addition to specific technologies like LongRun2, Transmeta is now considering licensing the designs of its microprocessor cores, similar to the business models of Arm and MIPS Technologies.

"We're looking at almost everything," said Greg Rose, Transmeta's director of marketing, in an interview at the Consumer Electronics Show in Las Vegas. The company has not said anything yet about the future of its microprocessor business, but it is very interested in expanding its licensing business, he said.

Moving to an intellectual-property business model is much less expensive than trying to sell processors, Rose said. Transmeta does not own and operate its own chip manufacturing plants, which is extremely expensive, but there are still significant testing and distribution costs that could be avoided if the company no longer sold chips, he said.

Transmeta entered the processor business 2000 with a low-power chip designed for ultraportable notebooks. The chip, known as Crusoe, consumed much less power than other notebook processors of the era but was unable to duplicate the performance of other chips.

Crusoe used a VLIW (very long instruction word) architecture, and needed code-morphing software to allow that technology to work with the x86 architecture used by processors from Intel and Advanced Micro Devices as well as Microsoft's Windows operating system.

This software-based architecture dramatically reduced the amount of power consumed by the chip because moved much of the processing from transistors to software. However, software can not match the performance of a transistor. Early reviews were unkind, and Crusoe's acceptance was further hampered in 2001 by manufacturing problems.

The company's latest processor, known as Efficeon, offers much better performance than Crusoe with the same power-friendly characteristics. However, Intel has since cornered the notebook market with the Pentium M processor, also designed for low power consumption.

Since its initial public offering in 2000, Transmeta has lost a total of US$591 million trying to get its processors accepted by the PC industry, as of its second-quarter SEC filing. The company continued to lose money in the third quarter of 2004, but licensing deals now account for over 50 percent of the company's revenue.

Transmeta will hold a conference call on January 21st to discuss the results of its evaluation, Rose said.

Follow Us

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.


Show Comments