Network: Whichever way the chips may fall...

Andy Grove, the chairman of the microprocessor giant Intel, is a man who is not afraid of controversy or scared of making hard choices. He has steered Intel through some difficult times, from a price crash in memory chips in the Seventies to the forced recall of early Pentium processors in 1994.

Dr Grove has a reputation for direct speaking. On his last visit to the UK, more than two years ago, he issued a wake-up call to businesses - adapt to the e-commerce revolution or die. Last week, his keynote address to the Confederation of British Industry was rather more muted. Dr Grove, whose most recent book is Only the Paranoid Survive, thinks Europe seems to be getting the message.

He joined Intel from the start, in 1968, and took over from founder, Gordon Moore, as its chief executive officer in 1987. Last year, he handed over day-to-day running of the company to Craig Barrett, the new CEO. But Grove retains the chairman's title and still plays an active role in shaping Intel's business strategy.

Grove's tour of Europe - taking in Dublin, Paris and Munich as well as London - comes at what could be another difficult time for Intel, which faces what could be its most serious competition in desktop microprocessors since the launch of the Pentium, in the shape of AMD's Athlon chip. Intel's next generation 64-bit processor, codenamed Merced, has suffered delays, costing it sales in the lucrative, high-end server and workstation markets.

The company is having to adapt to a world where the fastest growing market is for low-cost PCs. In the United States, PC prices are heading inexorably towards the sub-$500 price point. Grove himself believes the $400 PC is a realistic prospect.

Making the microprocessors that power these cheaper machines is a high- volume, low-margin business. And the PC market itself is coming under threat from a new generation of computing devices - handheld digital assistants, mobile phones and set-top boxes - that are not natural homes for power- hungry Pentium chips.

Grove, though, relishes these challenges. He is a firm believer in strategic inflection points for businesses, where new technologies shift the boundaries for everyone. "Companies at that point either adapt to the new construct and move on to greater heights, or they miss the moment and go into a decline," Grove said.

Grove reiterated his determination to compete with anyone who wants to enter the microprocessor business. There is no part of the microprocessor industry Intel is not interested in, he suggested, whether it is low-power chip sets for mobile phones or 64-bit technologies for the most powerful Web and e-commerce servers. Intel is already the world's leading maker of flash memory for digital mobile phones, and Merced will give companies such as Sun Microsystems and IBM pause for thought.

To see Intel as simply a component maker for desktop computers is to seriously underestimate the breadth of its expertise. But Grove dismissed accusations that Intel has an unhealthy domination of the PC market: "We have driven the development of microprocessors, but we have not had any lack of competition in 27 years," Grove said. "We feel the business is as competitive as it ever has been."

Intel remains very much a research-based company. Both Grove and Barrett hold doctorates and have published academic texts. Intel is not only investing in its own technology; it is also putting money into outside firms that can either improve the performance of PCs, or create new applications for PC technology. The guiding principle of Moore's Law - microprocessor power doubles every 18 months, and costs halve - is still very much alive at Intel. Grove sees Intel's external investments as an important way to find applications for the processing power Moore's Law creates. So far, Intel has invested in some 225 companies, with holdings valued at $3.5bn.

The force driving the market for computers, Grove maintained, is the Internet. He thinks that soon we will no longer talk about e-commerce as a separate form of business; instead, it will be at the heart of every business. "In some years' time there will be no such thing as Internet companies because all companies will be using the Internet in their business and in their internal operations," he said. Companies will use the Internet in two ways: to broaden their reach, and to reduce their costs. The smartest firms will harness technology to do both.

The Internet is creating an enormous market for high-end computers and networking technology: the building blocks of e-commerce. Intel makes and sells networking equipment aimed at smaller businesses, but it is Intel-based servers that will provide the greatest opportunity, especially for machines based on 64-bit chips. IA-64, as Merced is also known, will let companies speed up searches of their customer records, have better, faster encryption and security, and better visualisation tools. "We still have most of the Internet ahead of us," Grove said. "We believe that 96 per cent of the estimated server capacity that will be needed by the year 2005 is still to be deployed."

Grove, though, is a pragmatist and he realises that not all companies have the capital resources to invest in the hardware they need for effective e-commerce. Here, Moore's Law causes its own problems. Companies, especially smaller ones, are reluctant to invest as what they buy is sure to depreciate in months, not years.

As Grove pointed out, a small print or file server and a fully secure, reliable e-commerce server share few similarities apart from the name. And for smaller firms, the expertise gap might be as difficult to breech as the gap in price. Intel's solution is to change the way it sells its hardware. The company is about to open its first commercial server farms. Here, clients will be able to rent space on fully serviced computers in fully secure environments. "It is our intention, if you like, to provide our microprocessors a bit at a time," Grove said.

The concept is based on the server farms that Intel has created for its own e-commerce purposes. Running server farms pits Intel head-to-head against some very powerful companies, especially Internet service providers and, increasingly, the phone companies.

But Grove maintains that the move is a way of increasing access to best- of-breed Intel hardware, rather than an attempt to make Intel a services business. "Our mission is to be a building block supplier for the Internet, as we are for the PC business. But we realise that not everybody is in a position to buy this technology," he said. "But we are pretty good at making stuff and pretty good at designing stuff. Intel, he insisted, will always be a manufacturer. "We have to design the right silicon for the right market at the right time."