IBM to Buy SAP? Why Not?

If IBM were to acquire companies founded by disgruntled employees who went on to create IT players large and small of their own, it would be a massive company of great wealth and power that had all of the other players except those who left Hewlett-Packard, Sun Microsystems, Oracle, Intel, Microsoft, and a few other industry giants. That company would be the dominant supplier in the data center, spanning from servers up through operating systems, middleware, and applications software, as well as services.

And that IBM–call it Even Bigger and Even Bluer–would have dominant and probably monopolistic power in each of those markets. Gene Amdahl would have never left IBM to create clone computers, knocking out Fujitsu as a credible systems and then server player; H. Ross Perot would have never left IBM to found Electronic Data Systems; IBM would have never decided to give up on applications software, and would have kept its MAPICS, DMAS, CMAS, and other MRP systems from the 1970s and 1980s and would have never let go of the German founders of SAP, who would be in charge of the IBM Application Group today.

I am not sure if this would be a good thing or a bad thing, since I am distrustful of large aggregations of power of any kind. But, the idea may get a test in reality, at least if you believe that SAP is looking to be acquired by IBM.

Hasso Plattner, the chief of the supervisory board that runs SAP and one of the company’s founders, gave an interview to FT Deutschland, the German edition of the British Financial Times, where he once again raised the prospect of SAP being acquired, and said there were only three companies that he thought were a fit for SAP.

“There are only three potential buyers: IBM, Microsoft and Google,” Plattner explained to the paper. “I don’t see anyone else. If shareholders think that a combination, and not independence, is better, then it will happen.”

Plattner is not just shooting the breeze, even if he is thinking out loud; he is one of the four founders of SAP who are still at the company, and he has a 12 percent stake in the application software juggernaut. And like other top managers at SAP, he is not afraid to consider the possibilities of an acquisition of SAP by another IT player. In fact, SAP’s execs seem to think it is their fiduciary duty to contemplate this from time to time, and to talk publicly about it.

In the summer of 2004, SAP let slip that it had actually contemplated a merger with Microsoft. The company did this not out of a sense of drama, but because the raging lawsuit between Oracle and its hostile takeover target, PeopleSoft, was about to leak documents that showed the two had seriously considered a merger. But, because of the complexity of the merger and the difficulty of convincing American and European regulators that this did not give the combined entity a huge amount of power in the corporate computing market caused the two to back off. Microsoft certainly had the cash to do the deal back then, and to acquire SAP today would be a lot more expensive since SAP has grown (along with its share price) and Microsoft has returned tens of billions of dollars to its shareholders and, while hugely profitable, its stock price has languished.

Talk of Google acquiring SAP is utter pishtosh. Nonsense. Stupidity. But, financially possible, oddly enough. In its latest annual results, SAP had $10.9 billion in sales and has a market capitalization of over $66 billion; the company has 36,700 employees and 33,200 customers. In the ridiculous world of Wall Street, Google, a company with $6 billion in sales in 2005 and 5,700 employees, has a market capitalization of $116 billion. And if Google bought SAP–most likely by issuing a gargantuan pile of stock since it only has $8.5 billion in cash–it would be akin to AOL buying TimeWarner a zillion years ago in the dot-com era. And, in the long run, it is hard to believe that SAP would not end up taking over Google, just like TimeWarner, the old media company, took over and wrote off its AOL investment. If Google thought software as a service was its future, an SAP buy might make sense, but knowing how to run a search engine and advertising display system with a few hundred thousand servers is one thing. Building such a complex to support the data processing operations of potentially millions of ERP software users would be another thing entirely. Which is why SAP has been very cautious when it comes to SaaS. And why Google, if it does anything, will embrace an open source ERP suite and provide that as a service, it it does anything at all.

An IBM acquisition makes more sense, and is a much better fit. Pity, then, that IBM spent all of its cash on share buybacks. From 1995 through 2005, IBM spent over $70 billion on share buybacks and dividends, with about 85 percent of that being for share buybacks. That cash, invested in some other means for a decade, plus current cash on hand would have given IBM enough dough to pay for SAP at list price today, and it could always issue more IBM shares or borrow to get the remainder of the money. (I am sure that Carlyle Group, which is headed by former IBM chairman and chief executive officer, Louis Gerstner, might want a piece of such action.) IBM can do the deal. There’s no doubt about it.

SAP is, at its heart, a lot like IBM. It is a very top-down, tightly managed company, which stands to reason since it was founded by former IBMers who were frustrated by Big Blue’s lack of interest in creating the first wave ERP applications on mainframes. (OK, just how stupid was that, right?) Just as IBM started delivering its System/38 machines at the end of the 1970s, Plattner and his buddies Dietmar Hopp, Klaus Tschira, and Henning Kagermann left IBM Germany to found Systeme Andwendungen Produkte in der Datenverarbeitung, or SAP for short. These founders, by the way, still hold roughly a third of the shares in the company, which makes them not only rich, but true stewards of the company. There is no R/1, incidentally. The R/2 suite was the first ERP suite–meaning that it was an integrated suite of tools, not just MRP software; that it could span many large businesses; and that, most importantly, it could support multinational companies with different language, currency, and accounting needs. R/2 ran only on IBM mainframes, and it quickly became the preferred platform for large German businesses. With the R/3 suite, which came out in 1993, SAP branched out and started to support Unix servers and a more distributed architecture. By the mid-1990s, large companies were unplugging mainframes and moving to Unix servers in droves, and only a few years later, Windows became a viable platform just as people started working about all the changes they needed to make to their software for the looming Y2K crisis. The desire to compete, the roaring economy, and the Y2K crisis created this huge wave of investment in enterprise software, and SAP quickly outran every other ERP software maker in the market. (SAP supported Digital VAXes and IBM AS/400s, but this was always a tiny portion of its business.)

Since it started selling off its application software businesses in the early 1990s–just before SAP started to get traction, ironically–IBM has been trying to position itself as the Switzerland of servers, operating systems, and middleware. Rather than compete with the ERP software suppliers, IBM wanted to be their best friends. But, alas, with the massive consolidation in the ERP software business, it is the software suppliers, not the platform suppliers, who hold the power these days. They can support–or not support–any platform they choose. By and large, the key software players have decided not to support IBM’s mainframe and OS/400 servers, and even those that do support these platforms usually lead with Unix or Windows alternatives. And they are excited about Linux. While I think IBM’s Switzerland strategy has been a key in the adoption of WebSphere and Lotus middleware and DB2 database sales, I think this strategy has run its course.

Let’s look at the numbers. IBM sells about $3 billion a year in DB2 database software and about $3 billion a year in WebSphere software. I think that about 60 percent of that maybe half of that combined $6 billion is on the mainframe and OS/400 platform (with maybe some of the WebSphere being on Linux or Unix machines tied very tightly to the mainframe or OS/400 server). These customers are not going anywhere whether or not IBM buys SAP. So that leaves another $2.5 billion or so in database and middleware software that might be in play. Customers who like DB2 and WebSphere are going to ask for it, and while some ERP players may balk and go to JBoss or Oracle Fusion or the Microsoft Windows Server System for the middleware stack, I think trading a few billion dollars in middleware losses for the largest installed ERP base is probably a good trade. Especially if you believe, as I do, that the database and middleware markets are going to experience a rapid revenue contraction in the coming years because of the Open Source Effect. As open source databases and middleware matures and is priced a lot lower than closed source products, there is no way IBM, Oracle, and Microsoft can hold the line on prices.

The upside of an SAP deal is huge. Right now, HP has over 50,000 installations of SAP worldwide, which is a huge base into which HP sells servers, storage, operating systems, and services. If IBM bought SAP and tuned the software to run well on its systems, operating systems, and middleware stack (probably a merged IBM WebSphere and SAP NetWeaver stack), it could radically boost its market share among SAP customers. And I have a hard time believing those Unix shops that love Oracle or Windows shops that love SQL Server running on IBM’s System p and System x servers will suddenly, because IBM owns SAP, stop buying such machines. Why would they possibly care?

What they might care about is this: A tightly integrated stack of systems and applications software that takes the tight integration of the AS/400 architecture one final step further up the stack. In fact, IBM could even rebrand these future, bundled systems as “Application Systems,” bringing the AS/400 name full circle. And these future Application Systems would run, if customers want, other ERP suites, just like the iSeries supports AIX and Linux and has Windows co-processors.

The ERP market itself is not the money. The drag-along effect on servers, software, and services is. If you assume that about 60 percent of IBM’s total $91.1 billion sales is on mainframe and OS/400 servers, that is about $55 billion in servers, software, and services that is not going to change downward if IBM buys SAP. That leaves $36 billion or so in play. But, SAP itself will probably bring in at least $3 billion more in software sales in 2006, plus at least another $3 billion in maintenance and maybe $3 billion in consulting, so you are really talking about needing to fill in a $27 billion gap. How much of that $27 billion is for generic servers, software, and services that have nothing to do with ERP systems? Probably a lot. So maybe we are talking about something on the order of $10 billion to $15 billion in sales are really at stake. If IBM can’t drive at least $3 in server, software, and services sales for every $1 in SAP revenues on software, then the deal is a bad idea. But, you see, SAP already does that itself. And IBM, the master of services, can surely do better. (Maybe it will even buy EDS, eh? Get the band back together.)

The question you need to ask is this: How many more SAP installations can there be in a world where IBM has 500,000 enterprise customers, which is the largest installed base of enterprise customers in the world?

Which is why SAP might give some serious consideration to doing the following: Buying IBM.

RELATED STORIES

SAP Focuses on Web Services, SOA with mySAP ERP 2005

HP Boasts of 50,000 SAP Installations, Solaris 10 on X64 Gets SAP Support

SAP-Microsoft Collaboration Produces “Duet”

Microsoft, SAP Considered Mega Merger