With Hewlett-Packard Co heading for $18,800m turnover this year and Fujitsu Ltd at around the $26,000m mark, while Digital Equipment Corp looks hopefully to Alpha to rocket it off its $14,000m-a-year launchpad, the contenders to take over leadership of the mainstream computer industry from IBM Corp are lining up – and a major new round of mergers and acquisitions could be on the way. Until the last couple of years, Fujitsu looked the most likely company to catch IBM up in terms of size and reach, but the company is now beset by all the problems that assail IBM, led by an over-dependence on mainframes. In the mid-range, its key products are derivative, coming from its ICL Plc affiliate and Sun Microsystems Inc, while in personal computers it is an also-ran in Japan, not visible in the US, and only a significant player in Europe by proxy via ICL. Its strongest American card is its 44% stake in Amdahl Corp, and while IBM appears to be playing into Amdahl’s hands in the short term, Amdahl knows it has to build a big new business almost from scratch out of Unix and its Huron applications development and delivery environment if it is to grow much beyond its present $2,500m or so annual business. Nor is Fujitsu in any position to improve its position with another acquisition – its balance sheet is too stretched and it can’t use its shares – too few US investors would want to hold them.

Loser

Hewlett-Packard currently looks by far the best placed of the biggest manufacturers to come out as the winner in the 1990s but then three or four years ago, that was being said about DEC. At present, everything that Hewlett-Packard is doing looks right: it has won itself a reputation of taking better care of its customers, an example being the way it is not forcing its MPE users to move to Unix, instead letting them migrate at their own pace. In Unix, it has succeeded in creating the impression that it has far more of the management, security and support capabilities in place than any of its competitors, and while this may not be entirely valid, time and again in the computer industry, winning the perception is nine tenths of the battle: in the 1970s, it was universally acknowledged among the cognoscenti that Burroughs Corp had by far the best mainframe architecture and operating software, but that did nothing to prevent the vast majority of its users migrating to IBM over time. IBM was perceived as the winner, just as today, even the things it does right are in trouble because IBM is perceived as the industry’s most emphatic loser, and more and more people are talking seriously about the company being in danger of going the same way as Prime Computer Inc, Wang Laboratories Inc and Control Data Corp. And Hewlett-Packard, with its laser printers, has demonstrated that it still has the ability to build a major business from scratch, and its OEM disk drive business, although not very visible, is treated with respect by its competitors. Because all of DEC’s business is computer-related, it remains slightly bigger than Hewlett-Packard in computers – Hewlett has its medical electronics and instruments businesses as well, but computers are becoming more and more dominant with every year. The case of DEC is tantalising: either the Alpha will do for the company what the VAX 8600 – quickly supplanted by the VAX 8650 did for it in the 1980s, and the vast installed base of VAX machines will be replaced by Alpha AXPs, giving the company three years of fantastic growth that propel it to $25,000m, or the desperately late Alpha and the new OpenVMS will turn out to have too many bugs and glitches to be trusted, in which case, DEC will be trotting off after IBM.

Truly viable

There is an assumption behind all this that size does matter, which seems to be belied by the fact that IBM’s enormous $60,000m a year bulk has been unable to save the company – but IBM’s problems all arise from the fact that the company’s business is so hopelessly slewed to one sector of the market, with products that fewer and fewer people are going to be happy to own. IBM’s pers

onal computers business may be enormous, but it has never made very much money, and it is now more like a millstone than an asset. The company has only two attractive businesses now, the AS/400 at $14,000m or so a year, and the RS/6000, at barely $2,000m – which means that in the rankings above, the truly viable computer businesses of IBM lie between Hewlett-Packard and DEC. Economies of scale mean that size does still matter: to succeed, companies have to grow while constantly guarding against building the sclerotic bureaucracy that bolsters a sense of infallibility that was IBM’s ultimate downfall. And Helwett-Packard has shown itself willing to make regular infilling acquisitions as it has grown – the company was heavily castigated for the price it paid for Apollo Computer Inc, but without Apollo, it would not be in anything like such a strong position today. The two other big diversified computer companies are Unisys Corp at $8,400m or so, and NCR Corp at $7,100m – only the same size as Apple Computer Inc, a pure personal computer play with few designs on the data centre – and if AT&T Co is really serious about making it big in the computer industry, it will soon have to start thinking of buying NCR a present – and Unisys begins to look tempting now that James Unruh has finally go the company onto an even keel and Unisys’ own mainframe millstones under control. NCR and Unisys are both following the same Intel Corp with everything Unix strategy, and Unisys’ strong position in banking and reservations would provide new outlets for NCR’s strengths in special purpose terminals. Shareholders in Unisys have been through so much misery, the fear that the IBM mainframe blight will soon infect the company is so great that no reasonable offer for the company is likely to be refused – and while the Unisys debt burden is now manageable, it is still onerous for a company of Unisys’ size, but AT&T’s credit remains almost as good as gold. The main factor that would make AT&T look twice is that it would not likely play well with its own shareholders in the short term.

Crazy

Compagnie des Machines Bull SA is in such a bad way that only the determination of the French state keeps it alive, and Ing C Olivetti & Co SpA is not much healthier. One company that could still make an impact on the world computer market is Siemens-Nixdorf Informationssysteme AG – but only if Siemens is serious about remaining – outside Germany, becoming – a force in the computer industry. Siemens Nixdorf is in as bad a way as IBM, and only a couple of years behind Bull, and that simply because the German public sector still hasn’t realised it’s crazy to pay $50,000 or more per MIPS for computing power. The depth of the recession facing Germany means that the boom must fall soon, and then about the only way for Siemens to save the company will be to make a major acquisition that finally puts it onto the world stage. Companies like Sequent Computer Systems Corp and Pyramid Technology Corp are technologically compatible – but painfully small to make much difference, which suggests that a bold move for another technologically compatible company – with a nice sideline fit in telecommunications – Tandem Computers Inc, is the smallest acquisition that would make a real difference – but it’s hard to see cautious Siemens being that bold. Welcome to the New World Order.