Republic central to Intel's vision of tiny chips and king-size profits

Intel's Craig Barrett dismisses the recent scrapping of two top design projects as just a missed step in the tech giant's onward march, writes Jamie Smyth, Technology Reporter.

When Craig Barrett jumped on the corporate ladder at Intel in 1974, microchips contained 1,000 transistors and computers were capable of performing only simple tasks.

Thirty years on and the world has changed dramatically for consumers and corporations, in part driven by Intel's relentless drive to improve its technology, which acts as the brains of computers and electronic gadgets.

Intel's latest generation of microprocessors squeezes more than a million transistors onto a thin sliver of silicon, no bigger than a 10 cent coin. And computers - which in 1974 were huge machines large enough to fill whole rooms - can now fit comfortably into a top pocket.

By cramming more and more transistors onto a single chip, Intel has revolutionised the computing world for everyone.

But a recent decision by Intel to scrap two of its most sophisticated chip-design projects - Tejas and Jayhawk - is causing people to question the firm's ability to continue to increase the speed of its chips by cramming more transistors on a single slice of silicon.

A New York Times article last month described how Intel's newest microprocessor was running slower and hotter than its predecessor. To overcome the problem, Intel has been forced to adopt a strategy already pursued by its rivals: obtaining more computing power by stamping multiple processors on a single chip rather than straining to increase the speed of a single processor.

The change of course is a blow to Intel and could undermine "Moore's Law" - the theory that Intel co-founder Gordon Moore coined more than three decades ago that microprocessor speeds would double every two years.

In an interview to mark the official introduction of Intel's $2 billion (€1.7 billion) plant in Leixlip, Co Kildare, Mr Barrett - a materials expert and academic turned manager of 70,000 staff - bristles at the mention of the New York Times article.

"I was frankly very disappointed in the article," he says. "The net of the article was that we are effectively moving away from Tejas architecture towards a dual core. You could look at this negatively or positively."

Mr Barrett portrays this change of course at Intel as a small blip for a firm that can afford to spend heavily on research and multiple chip architectures. But this week new media reports suggested AMD would launch dual core chips in 2005 ahead of Intel.

The former associate professor of materials at Stanford University and author of more than 40 technical papers, also dismisses any premature talk of the death of Moore's Law.

"Since I've been in the industry everyone has always suggested that, 10 to 15 years out, Moore's Law will come to a screeching halt. But then talented engineers come up with solutions to problems that us older, more conservative people think are insolvable."

Up to now Intel has consistently managed to reduce the size of transistors by using more sophisticated equipment. Its new plant, Fab 24, uses 90 nanometre technology - a nanometre is more than a 1,000 times smaller than a human hair - and it plans to introduce 65 nanometre technology at a brand new fabrication plant called Fab 24-2 within two years.

But there could be limits to how small Intel can go, Mr Barrett concedes, who says that at sizes of five nanometre, everything breaks apart.

Intel is now researching new areas such as nanotechnology and replacement materials for silicon to try to overcome the problem. Some of this work will be undertaken by Intel researchers at Trinity College's nanotechnology centre, which is being co-funded by Intel Ireland.

The results of the research will be fundamental to the technological and economic development of the world over future decades.

Mr Barrett is a firm believer in the benefits of this type of core research and development activity and Intel has invested heavily during the recent recession, which was the worst in the history of the computer industry.

He commends the Government's decision to prioritise research and notes that all the most successful national economies are spending in this area. "It is the genius of new ideas for products and services that helps an economy to move up the value add chain," says Mr Barrett, who oversaw the decision to set up a communications research division in Shannon in 2001.

Intel Ireland has also benefited from the corporation's decision to spend $4 billion over six years on two new state -of-the-art manufacturing plants during his tenure as chief executive.

"Fab 24 competed with every place on the face of the earth for the Intel investment... We scour the world on a continuous basis for engineers and sites."

Corporation taxes, education and stable Government policies are the key reasons that Ireland has benefited from continuous investment from Intel since 1989.

And despite rising costs in Ireland, Mr Barrett remains bullish about the opportunities for future Intel investment here. "We started operations from an autosales room in Palmerstown in the late 1980s and built motherboards. We are now manufacturing microprocessors with 90 nanometre processes and expect to move to new technologies every two years," he says.

Ireland's economy has benefited from Intel's success to the tune of €400 million every year. And Intel has remained on top of its own industry throughout this period. "We don't release industry figures but technology consultancies that follow us show Intel has between 80 and 85 per cent of the market and that has gone up in the last few years," says Mr Barrett, who denies complacency.

"We are paranoid all the time and take all the competitors seriously... The biggest worry I have is our ability to get our own products out on time. If you don't have new products you die."

Intel's aggressive stance and its overwhelming dominance in the microprocessor industry has been noted by regulators. Japanese officials raided the firm's offices there recently and the European Commission has restarted an inquiry into alleged anti-competitive sales practices.

Mr Barrett says the firm has no case to answer in the inquiry, which has been restarted after a new complaint made by AMD.

"We recognised over a decade ago that our market position required us to behave differently from other companies and we have taken this into account with all our sales, employee and management training... I am comfortable that our policies and activities are all above board."

Mr Barrett, who is due to retire as chief executive of Intel next May, says he has made few solid plans for his retirement. Perhaps some fly fishing in Ireland and a rest after 50 years at the cutting edge of the technology industry.

But he says wants to stay in touch with the industry and has already accepted a part-time job with the National Academy of Engineers in the US.

He will probably still be close to Intel. After all, Mr Barrett jokes: "I intend to remain a shareholder in 15 years."