Unmitigated disaster in the Unix world faces IBM if a new report on the company is to be believed. As far as we can tell, the scenario the report – from International Technology Group, Los Altos, California – projects is not presented as a disaster, but that is what it would be. Because the group forecasts that the forthcoming RTs will outsell the 9370 in 1990 (that shouldn’t be difficult), but will also be outselling the AS/400 by 1993 or 1994 – disaster number one. Because that forecast is accompanied by a graphic showing IBM generating $4,100m from Unix systems in 1993. That’s not bad, is it – a four billion dollar business from zilch in only four years. But that $4,100m should represent only about 5% or 6% of IBM’s total business by 1993 – even less if the company manages 10% annual turnover growth overall. In other words, by 1993, IBM’s high margin AS/400 business will be well on the way to being wiped out by what by definition will be low quality, very low margin business.
Crippled
And that $4,100m is not concentrated in a single market sector, where there would at least be economies of scale. Instead it is split over four segments – AIX on the PS/2 accounting for $470m sales in 1993; RISC workstations representing $1,360m, 33.2% of the total IBM Unix business; Unix on RISC-based mid-range systems – multi-user machines competing with the AS/400 – only $1,220m, 29.8% of the Unix total; and Unix on large systems, $1,050m, 25.6% of the total. That last figure would be superb news for Amdahl Corp, which currently does only $300m out of Unix on its IBM-compatible mainframes. Actually, $300m is very good, but the growth rate of Unix systems for Amdahl Corp implied by the International Technology Group forecast is nothing less than phenomenal. Because Amdahl offers a native implementation of Unix for its 5990 machines where the best that IBM can offer is a crippled version that runs under VM, so that it is only likely to appeal to users with a small Unix requirement, who are happy to put the thing up in one partition of a PR/SM system. Amdahl would have to be the most hapless marketing company in the computer industry not to outsell IBM’s mainframe Unix systems two-to-one, which would mean that if IBM were to be doing $1,050m in mainframe Unix systems in 1993, Amdahl, currently a $2,000m-a-year company overall, would be doing $2,000m in Unix alone in 1993 – an annual growth rate of well over 100%. There’s no reason to suppose that it would not continue to grow its core MVS-based mainframe business at its current rate, which means that Amdahl is going to be a very substantial company indeed by 1993 – time to rush out and buy the company’s shares. The figures for RISC workstations for IBM are equally good news for the likes of Sun Microsystems, DEC and Hewlett-Packard Co, because according to the International Technology Group figures, IBM will still be doing less in workstations in 1993 than Sun Microsystems alone did in the year to June 30 last.
Pitiful
So on the projection, IBM is still going to be a pitiful also-ran in the workstation business in 1993 – and its strategy of confining Unix to a technical ghetto will have failed completely. Not that International Technology Group does not point in that direction in its qualitative analysis – it talks of a desk-top system in January 1990 at under $20,000 – but $5,000 is now the marker price for low-end Unix workstations ask DEC, Hewlett-Packard or Sun. Elsewhere the group looks for two tower models and a first mid-range model delivering up to 30 MIPS, with a 45 MIPS model following later in the year. AIX is being set up by IBM as a complete parallel environment to SAA, and will be coupled with Open Systems Interconnection networking to pose a real alternative for IBM users, says executive director at the group, Robert Simko. But on that scenario, that sentence should read to pose a real threat to IBM’s most profitable businesses. Is IBM really going to allow it to happen? The answer to that has to be not if the $1,500 report ‘IBM Strategies for the 1990s’ makes
it to Armonk and IBM’s top brass takes on board the forecasts contained in the Unix section. IBM has had to pay lip-service to a commitment to Unix because workstations are a major new market, and so much government business around the world is now specifying the Unix operating system, and IBM can’t turn its back on that business with equanimity. It has to have a Unix capability, and it has to find ways of pricing its bids so that they are competitive with those from the true believers while still making money for the company.
Intolerable
But if it allows IBM Unix to make big inroads into the System 36 base, in five years’ time it will again be faced with the intolerable situation from which the AS/400 and Systems Application Architecture were conceived to liberate it – two completely incompatible product streams becoming increasingly expensive to maintain and develop at the rate the market demanded. Why should the company willingly recreate such an unattractive state of affairs – especially when this time around, one of those two product streams will be at best marginally profitable in IBM terms. Because make no mistake about it, DEC and Hewlett have bitten the bullet and recognised that to remain successful, they have to get right down there in the gutter and mix it with the dirty-fighting discounters that will do anything to make a quick buck, and are having to rely on achieving enormously increased volumes to keep their margins up – and that is a game that IBM still shows no sign of being willing to play: its PS/2s are still hideously expensive by comparison with a string of reputable brands of high-end personal computers and this week’s price-trimming exercise and new models in the US were no more than an earnest that IBM is not prepared to sacrifice margins for the sake of putting the discounters out of business. And Armonk will allow the Unix scenario outlined by International Technology Group, with the ravaging of its mid-range systems base that it implies, come to fruition only over its collective dead body.
