IBM vs PSI: Goliath slays David

 

By Jeff Gould, CEO & Director of Research,
Peerstone Research

Some months ago I blogged about IBM's legal war with tiny Platform Solutions Inc. (aka PSI), which at the time was virtually the last company in the world still trying to sell IBM-compatible mainframes in competition with Big Blue.

The feisty little Silicon Valley startup had developed an ingenious firmware emulation of mainframe architecture that enabled it to run IBM's z/OS operating system on servers built with Intel's Itanium chips. Customers who bought PSI's plug compatible hardware would still need to fork over a boatload of money to IBM for the operating system and middleware, but it seemed like an arrangement that benefited everybody, because it gave smaller mainframe users a reason to stay on the platform rather than migrate to Linux or Windows. It wasn't long however before a noisily buzzing fly plopped into the ointment. PSI had naively believed that IBM would honor the promise posted on its web site "to grant nonexclusive licenses under reasonable and nondiscriminatory terms and conditions" to competitors who played by the rules. But IBM's lawyers had different ideas. In 2006 they took down the promise from the web site and filed suit against PSI, admitting with disarming candor in the court papers that: "IBM has refused to agree to license its patents to PSI and its copyrighted operating systems and other software for use on PSI's emulator systems despite PSI's demands because, among other reasons, IBM has no obligation to do so..."

Well, this sad little story has now come to an end. PSI is no more. IBM purchased the company this week for an undisclosed amount rumored to be in the hundreds of millions of dollars, thus extinguishing PSI's legal claims against it, and – more importantly – cementing a monopoly that is now more complete than at any time since Amdahl launched the first plug-compatible mainframe in 1975.

But aren't antitrust laws supposed to prevent this kind of thing? Surprisingly enough, the answer might be no. According to the Wall Street Journal, IBM believes PSI's revenues are too small to trigger action by either the U.S. Department of Justice or the European Union in Brussels. In other words, IBM is saying that since they already own 99.9% of the market, squashing the one remaining competitor who has 0.1% is no big deal.

There is still some hand-wringing going on in Brussels about competition in the mainframe market, and it remains to be seen whether IBM will actually get away with this audacious coup. Personally, though, I'm not particularly scandalized by the rude fact that IBM wants to eliminate its competition. I mean, how big a surprise is this? Economic theory tells us that every capitalist wants to do exactly the same thing (and so does every socialist, though you'll have a harder time getting them to admit it).

There is no need to wonder why IBM wants to protect its mainframe franchise from price competition. In a once fabulously lucrative market now suffering from long term and probably irreversible decline, the dominant vendor will naturally want to milk its position for as long as possible. If the market was still young, the presence of credible competitors would undoubtedly promote faster growth. But in a shrinking market even weak competition has the potential to spark a race to the bottom in pricing. Hence PSI had to die.

But is there really enough money left in the mainframe market to justify such blatant maneuvering? After all, IBM's mainframe revenues are only a fraction of what they were 20 years ago. They are now far smaller than Big Blue's huge professional services business, and they contribute far less to the company's bottom line than its modern software lines such as WebSphere, Lotus, Rational, Tivoli and now Cognos.

Indeed, the seeming disproportion between the relatively marginal place the mainframe occupies in IBM's total revenue pie and the amount of strategic attention the company still accords it suggests that we are looking at the problem in the wrong way. IBM has never been a product-focused company like rivals such as HP, Microsoft or Oracle. Rather, since the dawn of enterprise computing in the 1950s – a market which IBM invented – the company's strategy has been to install itself in the world's largest organizations and seek top-down control over these accounts' IT spending. Go to any large IBM mainframe user in the world – be it a bank, a government agency, a telephone company or a sprawling multinational – and you will find IBM employees permanently based on site, supervised by an IBM account manager who has nurtured a close and long-standing relationship with the customer's CXO-level decision makers. It is this strategic approach to account control that has allowed IBM not merely to weather the relative decline of the mainframe, but to leverage its still overwhelming dominance in that mature market into surging revenue growth in the adjacent non-mainframe software and services markets.

Using your overwhelming strength in one place to gain advantage somewhere else is a perfectly logical move – so long as the antitrust authorities don't notice, that is. In the 1950s IBM invented the concept of bundling software with hardware and became one of the most profitable companies in the world, but was later forced to unbundle by the Department of Justice. In the 1990s Microsoft built the Office franchise on top of Windows, and got similar attention from the regulators. Today Google is leveraging its strength in search to dominate online advertising.

The mainframe is no longer a growth market, but its central role in many of the world's largest IT shops makes it an invaluable lever in IBM's strategy to stay on top of the enterprise computing dog pile. I suspect that Big Blue's management team is crossing its fingers right now and hoping that the world's antitrust authorities fail to grasp the full scope of its strategy.