Big Brother comes to town

The IDA's most risk-laden gamble to date is its £16m backing for a now struggling multiinational, Trilogy, started from scratch to build in Dublin the most advanced computer technology yet seen. ALAN MURDOCH describes Gene Amdahl, the pioneering scientist behind the project, his revolution in micro-electronics, and the company's international strategy for tax avoidance. By Alan Murdoch

WHEN THE COALITION CABINET TOOK their seats for a meeting in October 1981 they were presented with a project that was by its proposer's admission - a minefield of financial, legal and technical hazards.

The company involved was Trilogy, a brand new Ameriican venture attempting at massive cost something two other US corporations had tried and failed - a new generation of micro-electronic circuitry intended to lead to the most powerful and efficient 'big brother' computer yet conceived.

What won over the IDA and then the Government was primarily, according to those involved, the record of the man leading the venture, the acknowledged leader in commputer design, the till-then highly successful ScandinaviannAmerican Gene Amdahl. There were reportedly few illuusions about the scale of the risk and the technical advances necessary before the plan could be implemented and a prooduct launched. But such were. the attractions - a share for

Ireland of a 'mainframe' computer market worth some twenty-five billion dollars a year by 1986, plus the training of young Irish graduates in the most advanced computer design skills around - that the risks were deemed worth taking.

The IDA finally committed £16 million to the Trilogy scheme (a £lOm loan, a £5m grant, and a £lm share purrchase), though details of the deal were not made public for two years. Before making the pledge, the agency spent several months contacting other overseas backers funding Trilogy, including Cii-Honeywell-Bull in Paris, and a consorrtium of US investors including the Bank of America.

Amdahl also expected to employ some 600 Irish workers from assembly staff up to production managers. Having dealt with the man for some six years before talks on Trilogy began in late 1980 (his previous operation, Amdahl Corpn, has a successful record here), there was already considerable trust in his ability to fulfil promises.

For Trilogy, Ireland offered minimal company tax rates, access to EEC markets, cheaper labour costs than in other European locations, yet a relatively high calibre of educaation among key workers. And topping all of this was the cherry of bigger state finance for the project than available in any of the other countries said to be pursuing the project. (Scottish and Italian locations were reportedly also disscussed.)

In the four years between the foundation of the commpany and the decision last month to put on ice the Trilogy computer, over two hundred and fifty million dollars went into investment in research and development at the company's Californian Silicon Valley headquarters at Cupertino. It was garnered mainly from major computer firms Digital, Cii HB, Sperry and Control Data. Their prime interest was in securing, in return for investment support, licensed access to Trilogy's unique and sophisticated new technology for incorporation into their own products. As these deals grew it became apparent that Amdahl anticipaated this part of the operation - supply of components to outside firms - yielding a hefty share of profits. This exxplains the continuing optimism about the company both from the IDA and Cupertino, in spite of the demise of the original end product.

FOR CLOSE ON THIRTY YEARS DR GENE Amdahl has designed the most powerful and profitable computers in the world. In the electronics world he holds - or did till last month - a status approaching Isaac Newton's in physics (the main difference is that Newton was lauded less in his own lifetime). Prior to his present ill-starred venture, he was the Daddy of the Big Brothers made by IBM in the 1960s, and went on to set up a highly successful multinational bearing his name from scratch.

Coming from a family of second generation Scandinavian immigrants (three grandparents were Norwegian and one Swedish) who settled as homesteaders in Dakota, Amdahl's technical brain showed itself at an early stage. At 12 he was constructing a counter-rotating propellor system for a heliicopter. He entered electronics during World War II, enrollling as a young recruit at a US Navy College. On completing a course in radio technology, he stayed on to teach the subject. He then took a physics degree, and in doing so helped build in Wisconsin one of the earliest modern commputers. (The first was probably that made by British miliitary intelligence to speed the deciphering of intercepted German communications.)

Soon after he joined IBM, and with them designed a string of successful systems that helped bring the computer into a wide spectrum of everyday commercial and governnment activities. They included the IBM 360, an enormously popular machine and a major profits earner for the commpany.

But while associated with major corporations all his working life, Gene Amdahl's commercial instincts have been anathema to their directors, particularly those at IBM. Aware of their exploitation of international distribution and marketing muscle in big profit margins, he has sought in his last and present ventures to drastically cut the costs of computing - mainly at IBM's expense. His own commpanies have vigorously pursued profits too, but there innoovation has received somewhat greater emphasis.

His departure from IBM - in fact his second leavetaking - he was one of the few people re-admitted to loyalty connscious IBM - came when directors rejected his plan for a highly sophisticated new computer, and closed the entire research operation behind it. Exasperated, he wrote a resiggnation letter that ignored the usual niceties. It announced he was going into direct competition with the giant.

An engineer rather than a marketing expert, Amdahl realised there were commercial advantages in concentrating on large-scale mainframe computers. There are relatively few customers for this type of machine, and the actual buyers tend to be informed specialists, so performance matters more than extravagant promotion. As they often have their own maintenance staffs, and tend to buy outtright or lease from third parties, less is required in backing for costly rental schemes - both saving the maker large sums. It seemed to Amdahl a clear opportunity to undercut a global giant in one of its biggest markets. He was already confident he could beat IBM on performance.

At first the new venture, Amdahl Corporation, struggled to find its feet. Its principal backers were the Japanese group Fujitsu, themselves manufacturers of large compuuters, and a Chicago venture capital house. Management failures and the apparent wish of the Japanese to oust Amdahl from control of his own company led to protraccted internal conflicts. Between 1971 and 1974 funds were running short, and Amdahl eventually saw his share diluted to a mere three per cent of the equity.

In contrast the computer was acclaimed by its first users as a remarkable feat of engineering. Based on dramatically condensed micro-circuitry coupled with improved cooling techniques, it achieved performance levels far in excess of IBM's best machines. The counter attack came in early 1977 with that company slashing thirty per cent and more off prices and launchingnew equipment. Within three days Amdahl matched the price cuts and for good measure reevealed a new smaller machine two years in the making, with performance and price in keeping with larger models' standards.

Thereafter progress was buoyant. The Amdahl compuuters earned a reputation for reliability and sales grew. By this year it had topped 780 million dollars in annual worlddwide sales. In Canada it achieved the unthinkable, deposing IBM as market leader, and taking over fifty per cent of sales.

Following discussions with the aDA from 1974, Amdahl Corpn came to Ireland attracted by low wages and corporate taxes, as well as access to lucrative European and Commonnwealth markets. (EEC countries have resisted J apaneseemade computers in favour of locally made products. The belief in Brussels is that open Japanese access to these marrkets will decimate home manufacturers. The Japanese element in US assembled Amdahl products hindered their US sales prospects.) Fears that Irish output would not meet required standards disappeared with the first shipments in 1978. These factors and the hefty government support on offer helped tempt the subsequent Trilogy venture in its wake. .

In August 1979 Dr Amdahl quit the corporation named after him (he remained as consultant and chairman emeritus until August 1980) frustrated with Fujitsu's by then domi-" nant influence that blocked research into areas he thought justified. All along he had sensed they were more interested in exploiting his technology for their own products than in the full development of his companies. The electronics industry in Japan has suffered from a chronic inability to conceive new technologies of its own, due in part to an authoritarian and highly pressured education system.

The American's answer to this defeat is Trilogy, a vehicle set up in 1980 to develop his most radical ideas on computer design, including the "wafer scale integration" of microocircuitry other companies had attempted but failed to bring off (see below). Amdahl commanded considerable loyalty and many colleagues left their safe jobs at the Amdahl Corpn to risk their futures in the new venture.

He was also joined there by his son Carl, then 29, a commputer designer like his father, who became vice chairman of the new company. Earlier this year, following repeated failures to meet design deadlines and rows between rival 'prima donna' scientists he stepped down. He is now a design consultant to the company.

Amdahl senior will be 61 this year, an age when most businessmen are turning their minds more to the golf course than to launching corporations. In spite of all the setbacks he still exudes enthusiasm: "There's a great sense of exciteement in it (the project). The motivation is basically in prooducing a technology superior to anything available from any other source."

Even 15 years after leaving IBM his animosity towards that organisation is still evident. During that time he has seen them take determined steps - propagandistic and technological - to block his advance. "They are good at what they do," he concedes, but lambasts them for their consistently monopolistic behaviour.

Amdahl has taken one of the distinctive features of IBM with him however - a deep mistrust of trade unions. He believes a company should resemble a family, that employer and employee should not be divided by other loyalties, and has intimated publicly that his new Irish plant will be nonnunion. Given the already conspicuously low level of unioniisation in Irish electronics, he seems likely to have his way.

Personally, he enjoys privacy outside the job, eschewing the customary high life of the Californian executive, preeferring golf, reading and the company of close family. Secluded from distractions, he applies intense concentration to intricacies of electronic design. Once, pacing up and down at home, lost in a mental maze of microcircuits, he fell down the stairs.

He is a quiet measured man with a cautious and deliberate conversational manner. Meeting this, interviewers suspect it would be easier to catch one of his computers off guard.

His wariness is stamped all over his ambiguous political endorsements. Though a registered Democrat, Amdahl believes his party are in no position to get their man in. He considers Reagan "has done a fairly successful job with the economy," and may even vote Republican this time. He denies the reason- is the, boost to demand for electronics further gargantuan defence budgets might provide. While that may be the case, he admits, "there are other things we need to do before we spend more on that." He accepts that high defence budgets draw heavily on money markets to the detriment of non-military sectors forced to pay crippling interest rates.

Despite current problems, associates put a brave face on the project "Dr Amdahl has great inner confidence in himmself and what he's doing," says one bf his collaborators, something that undoubtedly helps in winning over finannciers. Following one of the colourful seminars held in London, Dublin, and Edinburgh last September to woo them, a top insurance chief told the 'Observer': "This is undoubtedly a high risk investment ... but they impressed us as people who knew what they were about."

In excess of 250 million dollars has been put up in either share capital or loans. The company expects to reap reveenues running into hundreds of millions of dollars from allowing other electronics companies licensed access to its technology. (Current licensees are Digital, French based CiiiHoneywell Bull, Sperry.)

The decision to suspend the mainframe computer part of the Trilogy venture was influenced partly by the strong competition from IBM. Their Sierra computer, though not as powerful as that envisaged by Trilogy, is much closer to completion and is expected to appear next year. Amdahl told Magill that the arrival of a Trilogy machine would almost certainly lead to a repetition of IBM's ruthless rearrguard action of 1977. In a court case in the 1970s it was alleged by Fortune magazine in September 1977 that manufacturing costs of IBM's larger computers accounted' for only 15 per cent of the selling price. Even with margins only half that size, IBM would today still have the leeway to wage an equally ruthless price war.

Not surprisingly in such a hazardous business, Amdahl's associates are protective to the point of collusion. One describes him as "a very low key guy. If you were looking for a contrast you could say he was the opposite of John De Lorean." Another: "He's not a big corporate guy - he doesn't put across a big marketing spiel. He'd be the oppoosite of, say, John De Lorean."

It is not inconceivable that Gene Amdahl may yet get another chance to make a frontal challenge to IBM in its most profitable largescale computer market. But if that happens it is likely to be firmly under the scrutineering conntrol of a well-established corporate sponsor. The internal dissensions and the repeated failures to meet deadlines under his management make it doubtful any of the big rivals to IBM will again allow him full rein to run his own! show with their money. The value of disciplined manageement within the global giant all want to beat - IBM - is, ironically, the lesson the opposition are most likely to take from the Trilogy experience.

Footnote: one point Amdahl was never complimented on was his choice of company title. The dictionary definiition of 'Trilogy' is "a succession of three tragedies". •

******************

Trilogy - the super chip

To reach its TARGET OF A COMPUTING technology faster and cheaper than anything else on the market, Amdahl threw out several previously sacred conventions Of computer design. First to go was the notion that efficiency was inexxtricably tied up with smaller and smaller semi-conductors, or micro-chips.

He has spent tens of millions of dollars perfecting a new device with mul~iple lay~rs of il1icro-circuitry containing a h'uridredxfimesthe components'* found' in existing chips. With over two million on each wafer, the design was so complex the company had to invent new computer-guided engineering tools- to map out and etch the miniscule circuitry on to the 'super-chip's" surface. The assembled device is known as.a wafer-scale semiconductor.

All impression of the dimensions Dr Amdahl-is operating on can be glimpsed from the fact that - blown up to 1/10 inch scale - circuit diagrams of a single Trilogy semiconnductor. would cover an area bigger than Stephen's Green. It}s alsot1videJ11.,*in thewlact th~t mUell0f thedesign and testing'o(ihe ci)tnputer:~ functions can' only 5e done by other computers. .

The immense compression of circuitry is made possible by advances in the chilled-water cooling system linked to each chip, which helps in reducing overall computer size and drastically cuts power consumption. The The final result, it IS hoped will be a semiconductor WIth the same relation to the original microchip as the transistor to the radio valve.

The principal gain from all this is reliability. Having between two and three times the necessary number of cirrcuits on each wafer, it is possible simply to shift to a new circuit when another fails. This could be done automatiically by highly sophisticated programmes, capable both of pinpointing faults and activating the adjacent replacement circuits.

So confident has the company been ,of the performance of its device it has predicted after an average of one failure every 200 years in systems using it'as their baSIC component.

Manufacturing savings will also be made - unlike earlier chips the wafer scale conductors will function normally even with dozens of imperfections, eliminating expensive scrapping of rejects. .