If Transmeta's ambitions are fulfilled then the future will be awash with barefoot computer users, boldly going where none has gone before. The company's stated ambition is that "If it has a battery and a Web browser, it will be built with Crusoe".
Transmeta was established in the mid-90s with well pedigreed investment from the likes of Paul Allen (the co-founder of Micro-soft, whose Vulcan Ventures is a key financial backer), Soros Fund Man-agement, Deutsche Bank and Institut-ional Venture Part-ners. It features executive talent such as former Sun Microsystems engineer Dave Ditzel, the CEO of Transmeta, and Linux's inventor Linus Torvalds, now a key member of Transmeta's software team.
The company's intent was to develop a new class of computer chip conforming to what it describes as its four rules, which it believes will govern the future of computing. These are: first, that the Internet changes everything, including microproces-sors; second, computing is going mobile and so must microprocessors; third, full Internet compatibility is mandatory; and, finally, battery life and weight will dominate purchasing decisions.
The market into which the chip maker has launched itself is buoyant. According to IDC Australia, the local market for notebook computers has been growing at more than 13 per cent a year. At time of writing, the most current market statistics are for the third quarter of 1999 when, of the 457,928 personal computers sold in Australia, 74,982 - or 16 per cent - were notebooks.
IDC analyst Logan Ringland believes that within two to three years that figure will rise to as high as 25 per cent of the market. The firm has only just started to track the rise of the handheld computer, but Ringland believes that demand for these devices is also growing; he notes that the handheld computer is not a replacement for notebook systems, but a complementary device, pointing to even more opportunity for Transmeta.
Add to this the emergence of Wireless Application Protocol which will make portable Internet applications a reality, and it is clear why Ringland believes the potential opportunity for Transmeta to be significant, especially as the prices fall of other components, such as the displays, used in mobile computers. At present, he says, you could buy three desktop PCs for the price of a comparable power notebook, but Ringland predicts that this will change.
Transmeta will not have the market to itself, of course, as Intel is a very feisty competitor, and has recently introduced a new series of Pentiums for the mobile market. AMD is slated to follow suit later in the year.
What Transmeta brings, however, is a fresh approach. Unlike conventional microprocessors in which the instruction set is etched into the silicon, the Crusoe family has the instruction set entirely in software. The chip itself is therefore devoted to the core processing tasks. With a software, rather than silicon, instruction set, the number of transistors on the chip is reduced, so the power to run it is reduced, and the heat created by the processor is reduced. According to Transmeta, having the instruction set in software also means that it might in future be possible to simply download processor upgrades over the internet.
It has a remarkable potential.
And that is the lurking problem for Transmeta: it has to convert that potential into actual performance. It has to overcome the marketing prowess of Intel and AMD and the relationships which those companies have established with equipment manufacturers, in order to garner market share.
It is a problem similar to that which Linus Torvalds has already battled. With his Linux operating system he took on Microsoft, the world's marketing juggernaut. He proved that a good idea and grass roots support can combine to - if not topple - certainly jostle established giants.
With Crusoe, he may yet do it again.