Your browser is not supported. Please update it.

28 August 2003

Intel shifts wireless R&D focus from start-ups to universities

Given that Intel is now arguably the vendor providing the most powerful driving force for wireless, it is ironic that before this year it was facing a serious hiring problem in its wireless units, one that has created a new approach to acquiring talent that is likely to permanently affect the company’s product development strategies.

Intel’s wireless-obsessed chief technology officer Pat Gelsinger says that the reluctance of top flight mobile engineers to work for a ‘PC company’ with no wireless track record drove Intel to rely far more than it ever has before on university partnerships. This has proved so successful ‘ and so cost effective ‘ in key projects like the adaptable digital radio research that Gelsinger now tells Electronics Times that it will be a permanent shift away from full time hiring or acquisition of specialist companies, and towards academic partnership.

Gelsinger’s wireless group now issues 400 university grants a year, averaging $50,000. ‘I now have five of the top people on the planet working for me for $250,000. If I buy one company I have to at least spend a few million,’ he says.

He therefore believes that Intel will be less interested in acquiring start-ups for their talent and will focus on acquisition only for already proven technology advances ‘ as in its purchases of Xircom and the remains of smart antenna pioneer Iospan Wireless.

This does not affect the company’s policy of encouraging ‘ and keeping its eye on ‘ promising new technologies by investing in start-ups through the Intel Ventures unit, whose Wi-Fi fund is $150m. These investments show where Intel’s interests are lying ‘ from hotspot operators to digital signal processing (Intel invested in DSP specialist 3DSP last year).

Technology partnerships with established vendors are also an important way to accelerate Intel’s transformation from standalone PC chipmaker to wireless supremo. An important once is with Analog Devices (ADI) with which Intel has developed its Micro Signal Architecture (called BlackFin by ADI).

But for the more blue sky R&D, the universities have moved from being useful partners on the fringe of the Intel labs to being central to the whole effort, according to Gelsinger.

Gelsinger’s vision of the wireless future is now well known. Ubiquitous seamless wireless connectivity using the most appropriate signal and technology available at that moment ‘ cellular, Wi-Fi, WiMAX, UltraWideBand and so on. All this connectivity essentially free, with the industry deriving revenues from selling the devices and services on top, not from the connections themselves. And, of course, Intel providing the basis for all these forms of wireless networking, not with obscure customized technologies but in plain old CMos.

From its conception int 1998 until this spring, this vision was less public and Gelsinger was still encountering scepticism from within the senior ranks of Intel’s management ‘ which has embraced the Gelsinger risk with varying degrees of conviction. Some had a nasty suspicion that he was taking the famous words of Intel co-founder Robert Noyce ‘ ‘do not be encumbered by history: go off and do something wonderful’ ‘ rather too much to heart.

It was even harder to convince talented specialists to move from the start-ups of hi-tech boom California work in Oregon at the Intel R&D Labs, to choose a company with no RF experience and seen, in wireless terms, as irrelevant, rather than an RF giant such as Motorola or a start-up like Atheros. The all-CMos path deterred many too, since current thinking was centered on silicon germanium, gallium arsenide and other new generation chip technologies.

This drove Gelsinger and Pawlowski to the universities, and the oscillator at the heart of the first major demonstration of Intel’s CMos radio in June was largely the work of a student who now has a job at Intel. ‘The results are much better than what we would have got had we gone looking for the brightest and best RF guy,’ Pawlowski told EETimes.

Right up until Intel researchers carried out that presentation of their all-digital, all-CMos agile radio in June this year, the climate of opinion was that the chip giant had overstretched itself if it thought such as radio was possible, and that it was the company to deliver it. The demonstration started to warm that climate in Intel’s favour. ‘That’s when the world stood up and took notice,’ said project leader and Intel Fellow Steve Pawlowski.

This factor and the tighter employment market means that Intel is now inundated with CVs from wireless engineers and researchers. But although hiring has increased, Gelsinger says the new prominence of academic institutions in the R&D mix will be a permanent feature as the path towards the full implementation of the digital CMos agile radio, probably in 2006, unfolds.