British chips find life easier as french fries: Patrick Matthews on a Bristol invention with a 'Made in Grenoble' tag

Click to follow
The Independent Online
A BITTERSWEET little event at the Victoria & Albert Museum has gone virtually unreported by the British press. When the Franco-Italian semiconductor firm SGS-Thomson came to South Kensington last month to launch its newest and most complex microchip, it was also dotting the i's on the death certificate of Inmos, the firm that Britain's electronic dreams were made of in the Eighties.

SGS-Thomson's new T9000 transputer, a fast microprocessor carrying an unprecedented 3.3 million transistors on a single sliver of silicon, was invented at the Bristol design centre of Inmos, an entrepreneurial company set up with pounds 50m of public money in the late Seventies. But unlike its predecessors, the T9000 will be manufactured not in Britain, but at the Grenoble factory of SGS-Thomson, which has owned Inmos since 1989.

What was being laid to rest at the V&A was an extraordinary, almost romantic episode in the history of British electronics. Just as John Major's government has suddenly rediscovered the importance of manufacturing industry, so a decade and a half ago, when Inmos was set up, the dream was to help to reverse Britain's industrial decline.

That dream was given substance in 1978 when the Callaghan government set up Inmos. The chosen instruments were to be academic computer scientists, reborn as designers of the world's most advanced microprocessor. Iann Barron, chairman of Inmos until 1989, remembers how he selected his team. 'I just rang up all the major departments around and said, 'If you send your two best people, we'll interview them and see if there's a place.' We got the cream,' he says.

The young graduates could work the hours they chose - in one case, a regular 4pm to midnight shift - at their custom-made work-stations in premises outside Bristol. Their designs were turned into silicon in a gleaming Richard Rogers-designed factory on the outskirts of Newport, Gwent, opened in February 1983.

'The whole culture was different,' recalls one employee, who had come from Britain's mainstream, defence- orientated electronics industry. 'In my previous job, everyone was over 50; they were all very proprietorial about where they sat. Here, even now, the average age is about 30. It's just much more stimulating.'

Until 1983, Inmos had been essentially US-based, reflecting the original partnership between Mr Barron, a British academic, and Richard Petritz, a Silicon Valley scientist/entrepreneur. The initial investment by the National Enterprise Board went on setting up a factory in Colorado to make American-designed microchip memories. This US technology gave the newly recruited Newport workforce its first experience of making integrated circuits.

The transputer - the first all-British device, launched in 1985 - really was something new: the first microprocessor, with its own memory and communications links, designed specifically for 'parallel processing', in which a task is shared among an indefinite number of chips. It was novel in that it enabled chips to communicate directly with each other, and because it used a new language, Occam, purpose-made for parallel processing.

'We thought everybody would rush in and support it,' says Ian Pearson, now director of the Transputer Business Unit. 'But it was just too different for the majority of corporations to take on board.'

Inmos was an unloved child for which Mrs Thatcher's administration had neither time nor place. In 1984, the Government unloaded the company on to Thorn-EMI - a group progressively getting out of electronics and into buying the copyright on 'Singing in the Rain' and other popular songs - and Thorn sold it to SGS-Thomson in 1989.

These years of uncertainty meant that senior management at Inmos, rather than overseeing the production and marketing of the transputer, had to spend most of its time fighting the company's corner with an indifferent Department of Trade and Industry, then having to repeat the exercise with their new masters in private

industry.

After years of losses for Thorn-EMI, transputers are now making profits for SGS-Thomson: it sold 285,000 last year, worth more than pounds 20m, and its new version, 10 times more powerful than its predecessor, is expected to raise the value of sales correspondingly.

Two things have changed. One, according to Mr Pearson, has been the gradual cultivation of customers for low-cost, high-power processing: makers not just of supercomputers but, increasingly, of so-called 'embedded' applications (colour photocopiers, voice processors, handwriting recognition, personal navigation-by-satellite systems and microchip car controls).

And the other, the change of ownership, has given Inmos a new image. Customers now have reason to believe that, despite their novelty, the products will continue to be available, given the security offered by the world's 13th-largest chip-maker.

Overseas control, however, has had its costs - not least, the separation of British design and British manufacture. At Newport, the new owners swiftly reduced the product range - at present, it is running at only one- third of capacity - and introduced waves of redundancies. SGS-Thomson has invested nearer home by building a plant near Grenoble, run by a former Newport manager, and it intends to move all Inmos production there at the end of the year.

The various applications of the transputer seem to show that designing chips can cross-fertilise other hi- tech industries. Inmos itself has spawned related businesses in supercomputing, video telephones and, in Mr Barron's case, a company specialising in virtual reality.

But now, despite the British investment, neither British manufacturing industry nor the British taxpayer has any special call on the long-delayed transputer pay-off. And Inmos staff eagerly await the opening of Bristol international airport and the faster links this will offer with their parent company's main centres.

Mr Barron contrasts the benefits that he says Korea and Japan have gained from their governments' willingness to invest the massive sums required for high-volume production of microchips.

He is untroubled when accused of favouring a drain on the public purse. 'In the UK, we had a much better technological foundation than either Korea or Japan,' he says. 'If they can do it, why couldn't we? I think 'throwing money at the problem' could be just what you need.'

(Photograph omitted)

Comments