If you haven’t asked yourself why you didn’t buy IBM shares near the bottom in 1993, you are certainly in the minority. And if you own the stock now, particularly if you buy a lot of information processing goods and services, you might well be wondering whether it is near a peak… or approaching the end of a bull run. The IBM story during the past few years has been one of the most dramatic in business history. It was only a few years ago that the company seemed to be at death’s door, or at least on the steps of the bankruptcy court. But now, at least according to many of the financial world’s leading lights, IBM is not only sound of mind and body, but possibly as fit as it was in the 1960s and 1970s. (Even from the close perspective of the 1990s, it is pretty clear that IBM’s impressive profits in the 1980s were due more to non-recurring factors in tax policy and politics than to sustainable behavior on the company’s part.) The most remarkable aspect of this rebound in IBM is its basis.
Broken home
IBM is moving ahead not because it has gotten on top of the personal computer business. Compaq currently boasts that distinction. Nor is IBM the dominant supplier of Unix workstations or servers. The market is too fragmented for any one vendor to be considered dominant. IBM is at best a perennial contender for the lead in this low-margin segment of the industry. Nor is IBM at the forefront of the explosion in network-based information and communications services. The company’s Prodigy child is soon to be the offspring of a broken home; Sears has put its half of the business up for sale. America Online is the on-line information provider with the most customers in the US and the business is still in its infancy (which it might not outlive) in other countries. In the Internet game, the two US leaders will be AT&T and MCI Communications, the giant common carriers, while the overseas market is still up for grabs. No, IBM’s spectacular financial performance in 1995 is due mainly to the resurgence of the mainframe hardware and systems software markets. Another major component of IBM’s rebound in revenue (but not particularly in profit) is the company’s growth in full line information processing services. In addition, the proprietary AS/400 made a good contribution and, of equal importance, reached a stage that is very likely to yield excellent results during the coming year. However, the low end of the AS/400 market – the source of future large AS/400 sites – is under severe competitive pressure due to the battle between Windows NT and Unix server makers. IBM is one of the major combatants whose struggle for market share threatens the long- term growth of the AS/400 base. Despite IBM’s excellent results in 1995, there is certainly good reason to debate the ultimate fate of an IBM whose profits stem largely from MVS licenses and mainframe sales, and whose revenue has been enormously enhanced by a services business that will never provide the high profit margins characteristic of successful hardware or software product lines.
By Hesh Wiener
Moreover, there is a connection between IBM’s booming but not particularly profitable services business and the mainframe markets that have produced huge profits during 1995. A big customer who turns to IBM to provide enterprise services is very likely to be a shop with central systems based on IBM mainframes. Companies whose central computing needs are to be met by Unix systems or NT servers may not be as attracted to an IBM services offering because IBM has yet to demonstrate that it is at the leading edge of the open systems implementation field or the Internet-intranet services sector. Those young parts of the information processing industry remain the province of companies whose entire future depends on making their new technology offerings superior to the proven mainframe systems they hope to supplant. Thus, IBM is, for now at least, in conflict with itself. While it may have the financial and intellectual resources to become the pre-eminent supplier
of Unix central systems, it has a very strong incentive to keep MVS systems at the center of its strategy. Even as MVS is enhanced to provide Unix capabilities and obtains the right to be called a version of Unix, it is incontrovertible that MVS is not Unix in the most important sense. MVS is not an economical operating system running on economical hardware. Rather, it is a costly operating system running on systems that are coming down in price but that are still expensive compared to alternatives. Users that depend on IBM mainframes justify the cost of their machines on the basis of their legacy applications. During 1995, this led to a huge increase in the number of mainframe MIPS sold by IBM. The upturn in IBM’s mainframe business has been fueled by the punishing costs of using 3090s – yes, they are still out there in droves – that made migration to a CMOS 9672 extremely attractive.
Moved to CMOS
During 1995, around a third of the 3090 base moved to CMOS and this year at least half the remaining 3090 users will migrate. But at some point, IBM will have completed the harvest. It will become difficult for IBM to sell more mainframe MIPS except at much lower prices per MIPS. And that could lead to a downturn in mainframe hardware revenue. In 1995, IBM sold 59% more 390 MIPS than it did in 1994, and yet its 390 revenue rose by no more than 5%. Moreover, because such a large part of the mainframe market was based on replacement sales, the net increase in installed MIPS did not rise nearly as much as IBM’s sales results might suggest. And IBM did not increase its 390 software revenue proportionally to the net MIPS increase because of software discounts in its bundled leasing deals. In the unfolding scenario, IBM may have rescued the mainframe, and with it the investments of mainframe customers and its shareholders. But until it proves that it can prosper in other segments – notably personal computers and small servers – it may find that it is far easier to build revenue than to improve profits. This is a problem.
From the April 1996 edition of Infoperspectives published by Technology News of America Co Inc. All rights reserved.
