The European Commission's recent decision to launch two parallel antitrust probes of IBM's suspected abuse of the mainframe monopoly raises hopes that this huge market may open up in the not too distant future.
The mainframe business is a gigantic opportunity. For now, no one in that area can compete effectively with IBM, which leverages its mainframe monopoly to sell not only hardware but especially software and services.
The numbers and the strategic implications
IBM owns about 40% of the $24.5 billion market for mainframe software, which is roughly twice the size of the market for Linux-based software. In the total of hardware, software and services, IBM generates approximately 25% of its revenues and more than 40% of its corporate-wide profits in the "big iron" business. That's a profit figure somewhere north of $5 billion (easily), and it's anything but a business in decline.
Companies formerly foreclosed by IBM’s conduct will now want a piece of the action.
Those numbers reflect merely a part of the overall opportunity. Just the largest 10% of mainframe customers have collective annual revenues of $21 trillion, exceeding the GDP of the European Union and the United States.
Mainframes are only a part of their IT budgets. Those top 10% of mainframe customers have an estimated total annual IT budget between $800 billion and $1 trillion based on typical percentages in the relevant industries. If only half of that goes to external vendors and service providers (now or in the "cloudy" future), that's a $400-500 billion opportunity for the industry, or roughly 30% of global IT spending.
IBM does a lot of cross-selling (of Power CPU and Intel-based systems) to many of those customers, benefiting from the privilege position it owes to the mainframe monopoly. If and when the market opens up, IBM will lose its exclusive mainframe gatekeeper status and other vendors will compete more effectively for those accounts.
I've seen the slides of an IBM-internal presentation. "Account control" (in terms of controlling customers) is central to IBM's strategy, and the mainframe monopoly is the key to it. The 40%+ figure of IBM's corporate-wide profits doesn't even include the effects of that. That percentage relates just to sales of mainframe hardware, software and services, irrespective of upselling.
This is all the more important as the IT sector is undergoing a major transformation toward the cloud computing model. There hasn't been a similar need and (especially) opportunity to stake the claims in IT for quite some time.
In this recent posting I described how IBM's new mainframe generation -- the zEnterprise -- is designed from the ground up to "assimilate" (as a journalist put it) Intel-based platforms in the corporate data center. So what I previously called "upselling" (of non-mainframe offerings) is now about to become part of the "data center in a box", or "system of systems", that IBM calls its new mainframe.
It's a given that various IT industry giants will rush to secure a piece of the action, if they think that IBM’s abusive conduct to protect its monopoly will come to an end. They need the regulators to open up the market, but they won't wait until the end of the process. In anticipation of positive things to happen, everyone will try to secure the pole position through strategic investments.
The fact that the regulatory process has only just begun will, of course, be factored in when determining company valuations.
These types of situations and processes aren't unfamiliar to me. I co-founded one of the first online gaming startups in Europe, and we had strategic partnering inquiries from different telecommunications and media companies. One of those talks resulted in the acquisition of our company by Telefónica in early 2000. Subsequently I became involved with MySQL AB as an adviser to the CEO and early-stage shareholder in the company, and later saw the likes of Intel and SAP come in -- and finally, MySQL's acquisition by Sun.
I'd just like to describe from my vantage point after 25 years in the IT industry which players I would imagine to consider investment opportunities in connection with a future competitive mainframe market. The list below is ordered alphabetically. I'd like to clarify that at the time of publication of this posting, I do not own stock (or related derivatives) of any of the companies mentioned.
BMC: This $2 billion company does a large part of its business on the mainframe. When the cards get reshuffled, it won't want to be left behind.
Dell: This computer maker is more diversified than most people know and recently experienced an increase in profitability due to strong sales of enterprise hardware (server, storage and networking products).
Intel: Previously invested in mainframe emulation company Platform Solutions, Inc. (PSI). Can supply high-performance CPUs to power software emulators and could also play a role in hardware emulation.
HP: A natural IBM competitor with a strong foothold in the enterprise market.
Micro Focus: This company is also publicly traded and offers Visual COBOL, a .NET-based implementation of the programming language in which most mainframe legacy software is written. Emulators could run legacy programs on the same servers as Visual COBOL programs, making a gradual migration of workloads a more viable option for customers than it is today.
Microsoft: Invested in PSI (like Intel) and in T3 Technologies. For Microsoft's enterprise software division, improved interoperability with mainframe legacy workloads is essential. I'm not worried for free and open source software: if Linux couldn't compete in such a server-based context, it would never be able to compete with Windows. Customers should have all options. Compared to IBM's monopoly, competitive pressure from Microsoft and its partners would definitely be an improvement. By definition, there can never be two monopolies in the same markets.
Oracle: In its core business (relational database management systems), Oracle misses most of the mainframe opportunity due to IBM's stranglehold on the market. The standard mainframe database is IBM DB2. In a more competitive market, Oracle would be able to sell 11g and its enterprise software to a larger number of IBM's customers. Furthermore, Oracle's hardware division (formerly named Sun Microsystems) could play a key role in connection with emulation (similar to what I wrote above about Intel).
SAP: While not nearly as disadvantaged by IBM's practices as Oracle and others, SAP would also benefit from an open market. Its recent acquisition of Sybase is a cornerstone of its enterprise cloud computing strategy. The integration of mainframe legacy workloads with mobile and other cloud-related technologies will create new opportunities for SAP.
All of the companies I mentioned, and presumably a number I didn't even think of in this context, know that there is a lot of pent-up demand in the market for more competitively priced solutions for the execution of mainframe legacy workloads.
I wouldn't be surprised to see significant activity in the wake of the launch of antitrust probes. In my experience, such deals are often negotiated in a matter of weeks. Some take longer. But I can't see how the industry would forgo such an opportunity.
It's unhelpful that IBM always tries to use actual or suspected activities of other companies to deflect attention from the serious issues that need to be addressed in the course of the regulatory process. There's no denying the jugular importance of the mainframe to the world economy. IBM maintains its absolute control over this strategic platform with threats, intimidation, FUD and generally anticompetitive behavior, all of which stifles innovation.
Those are the problems. More competition is the solution.
So when the deals happen, let's let IBM cry. Its customers -- who are locked in and get overcharged -- have suffered long enough.
If you'd like to be updated on patent issues affecting free software and open source (including but not limited to the antitrust investigations against IBM), please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents.
