These days, the only computer that I personally own is an iPad. This happened because some years back, I developed an unhealthy fascination with unusual computers. This culminated with a purchase of a Digital AlphaServer 4100. It was a magnificent machine.
Unfortunately, as it turns out, it would also trip my electricity supply when I turned it on, and cost a small fortune to run, and only fitted in the kitchen, where it was universally unpopular. It had a ton of moving parts to go wrong, and I realized that my fascination with weird computers had to stop, so I donated it to the National Museum of Computing at Bletchley Park. I hope they took care of her. If you read this and work there, please let me know!
And so, one of the things that I realized at the time was that UNIX was dying – I realized this around the turn of the century. In the last 15 years, IBM has swallowed almost all of the UNIX market share and some vendors have simply left the market of been acquired: Digital, Silicon Graphics, NeXT, and yet others have lost massive market share: Sun/Oracle and HP/Compaq/Digital, to be specific. I called the HP Superdome as dead as a dodo in 2012 and if you think I was wrong, check this.
As for IBM, they have increased revenue in a declining market, which is in a very similar position to Pearson (the world’s largest book publisher), who are growing in a declining market. IBM’s UNIX hardware sales department is very profitable and ensures sell-on of even more profitable software products, and services to implement those software products. IBM likes high-margin and is an execution machine.
And so IBM sold off its low-margin PC business to Lenovo in 2005, which was an extremely smart move. Lenovo knows how to make good quality hardware at a good price and they are performing well, though at low margins: $584m operating income on $29.57bn of revenue in 2012. Both companies have prospered.
For the last few years, IBM and Lenovo have been dating, discussing the sale of its low-margin Intel Server business. From what I can see, IBM waited too long and only got $2.3bn, when they wanted much more – indeed they would have probably achieved double that, two years ago. The logic is – get rid of the low margin business and focus on the good stuff.
The technology market is a very tough place to be right now, and it’s very tough for a company like IBM to make good decisions. The biggest hardware players in the world now are the big cloud companies like Amazon and Google, both of which build their own custom data centers and don’t buy from IBM or HP.
But I question IBM’s decision. UNIX is a profitable market but a deceptive market. There are mission-critical use cases that use IBM’s System z and there will probably always be a market for certified, bulletproof systems. The thing that worries me is IBM’s middle ground – System p. System p is advocated by IBM as a low TCO UNIX environment, and that, to my mind, is a contradiction in terms. To add to that, IBM have combined System i (more mission critical than p, but not as much as z) with System p, to avoid the cost of developing yet another infrastructure.
Even worse, news outlets and IBM alike have referred to the Intel “System x” business as Low-End. SAP runs its entire business for 55,000 employees on one IBM System x3950 with 80 cores and 4TB of main memory. IBM sell a System x system with 4480 cores and 56TB of main memory. Don’t be fooled – System x starts at low end, but has systems that can solve the world’s hardest problems. I have no doubt that IBM didn’t want to sell this part, but Lenovo very smartly took it.
What I believe will happen is that companies will stop using HP Superdome, Oracle M-Series and IBM System p in the next 10-20 years, because they will move to the cloud – either in their own data centers, or in someone else’s cloud. There is no use case for “fairly” mission critical and whilst IBM is the clear leader in the UNIX market, it is a dying market.
So this leaves IBM without all of the Intellectual Property that made it – in my opinion – the finest Intel server manufacturer and with a declining server play. Lenovo will stop pimping IBM software and services in the medium term.
If there is one thing for which I have the utmost respect for IBM, it is its ability as a company to foster talent and reinvent itself and even when IBM was at the brink of bankruptcy – losing an impressive $16bn between 1991 and 1993, it managed to turn around its business and become one of the greatest technology companies in the world.
In the midst of this, IBM is focussing on its impressive Watson technology as a solution to part of this, whilst SAP HANA platform eats the DB2 database’s lunch. Note, I sell SAP HANA services, but the numbers back me up: HANA is growing in sales 80% year on year to just under $1bn in 2013, and IBM software grew 2.8% last year (they don’t break out DB2).
IBM has just posted its seventh consecutive quarter of decline. I dearly hope that Ginni Rometty has a master turn-around plan and IBM will once again rise from the ashes, because IBM is an incredible company that has survived where others have failed. But I worry that with all the other changes happening in the technology world whether selling off all IBM’s hardware business will be viewed as a mistake. We shall see.