Ditzel, who is travelling outside the country, was unavailable for comment.
"Basically, the [software emulation] effort started back in 1995, and our technique is different than Intel's architecture," said Ed McKernan, director of marketing at Transmeta.
Transmeta's Crusoe processor, which the company claims is capable of running complex operating environments such as Microsoft's Windows 2000 with as little as one watt of power, has only very recently begun to appear in ultra-thin portable computers from manufacturers such as Sony and Fujitsu. And McKernan feels that products from those manufacturers, combined with a strong market for Crusoe in the Asia-Pacific region, are early indicators of Crusoe's future US market.
"The workforce in Japan, for instance, is more mobile than the US workforce," McKernan said. "What's going on is that Crusoe is allowing the Japanese [manufacturers] to innovate more than they have before. And what we are seeing as trends from our Japanese [manufacturer] friends with the ultra-light mobile computers, it appears, is becoming the mainstream market."
As for Intel and AMD, McKernan said "the [processor] market is segmenting into two camps," which he identified as the low-power market, to which Crusoe is specifically targeted, and the speed market, which Intel and AMD occupy.
"Transmeta felt [it] had a different approach that was novel and unique enough and harkens back to the days of what RISC [reduced instruction set computers] stood for, a simple design that made for a low-cost, and in Transmeta's case a low-power, chip," said Richard Partridge, a senior research analyst at D.H. Brown Associates.
Partridge believes that Transmeta's premise was that advances in processor fabrication from companies such as AMD and Intel would eventually "drive processor clock rates so far beyond what people need" that the focus of development would shift from faster speeds to more efficiency, such as extended battery life through power conservation and system optimisation. At such a time, Transmeta would be there as a leader in processor efficiency, Partridge said.
"[Transmeta] knew that if [it] took a far different approach and extended battery life, then that would be something significant, rather than having a machine that runs a hundred times faster. They are betting technology will improve and that they will bring the power savings," Partridge said.
Intel made no official comment concerning Ditzel's remarks, but a company source suggested that the real reason Transmeta chose to develop its Crusoe processor from a software angle rather than along the guidelines of Intel's architecture was to avoid prosecution from Intel.
Unlike AMD, Transmeta does not have a cross-license agreement from Intel to build Pentium-style processors, the source said.
If Transmeta had designed Crusoe as a Pentium clone, "the only way Transmeta could have produced it would be in a protected foundry," said Dean McCarron, an analyst at Mercury Research.
A protected foundry would be a company that has a patent cross-license agreement with, in the case of Transmeta, Intel, McCarron explained, citing an incident in the late 1990s in which VIA Technologies Cyrix was sued by Intel over the Intel-like design of VIA Cyrix's M-2 processor.
A court threw out the Intel suit when it was discovered that VIA Cyrix was producing the chip in plants that did possess the proper Intel cross-licenses, namely IBM. and Texas Instruments, McCarron said.
"We went down a different path because we saw the advantage to the software approach," Transmeta's McKernan said.