IBM’s X86 Exit Strategy: Arguing The Good And Bad
February 3, 2014 Dan Burger
The news that IBM is selling its X86 business travelled like a flash of lightning. Then comes the thunder, which lasts longer and speaks louder. The thunder is the interpretation, the explanation, and the ramifications of the X86 decision. IBM, no doubt, puts considerable effort into making major business decisions like this one and then an equal amount of effort into making the decisions sound good to as many people as possible.
And why not? We’d all be happier if we saw only good in everything, right? What good can come from questioning the decisions of IBM executives? The cards have already been dealt. You have to play the hand you have.
But all the applauding and shouting “Bravo” as if this is just another in a long line of brilliant IBM strategic steps has little to do with business strategy or ecosystems. To walk around this deal and suggest that no matter how you look at it, this looks good, is to suggest there’s nothing to see here. To sum it up by saying the X86 market was never a good fit for IBM because IBM is a quality manufacturer not a quantity manufacturer is kicking the can down the road.
It goes beyond quality and quantity. It was simply difficult for IBM to navigate the X86 waters. Over the years, IBM has struggled at meeting the price points expected in this market, moving quickly through a very price competitive sales cycle, and ultimately delivering support after slicing prices to the bone. Competing in this arena did not fit the IBM business model. IBM’s evolving business strategy to de-emphasize hardware and concentrate on software and services is part of what’s happening here, but it doesn’t explain the entire picture.
If you concede that IBM was not good at the X86 game or that it is simply shedding a hardware business that is not part of the business plan, then getting out of that market might have been the best decision IBM could make.
But how can you sidestep the reality that at least 95 percent of the businesses on this planet are either using X86 servers or are considering them? That’s a big part of the market to ignore. That’s not to say IBM could tap that market any more successfully if it continued building its own X86 servers, but those servers–with the IBM brand–could be seen as a key to unlocking doors, a conversation starter that led to topics concerning software and services.
The plan, as explained by IBM in a conference call to its business partners last week, is that the sales channel will not notice a change in terms of product availability, customer service and maintenance. The only difference is that the X86 servers will no longer have the IBM logo glued on them. The transition, it was promised, will be seamless and robust. In other words, the key to those X86 doors will remain in friendly hands.
To make certain that happens, it appears IBM has sweetened the deal with Lenovo by offering a global OEM and reseller agreement for storage systems, including the Storwize disk storage systems built for the small to midsize market, and X86 software system that includes Systems Director systems management tools, its Platform Computing cluster management software, and its General Parallel File Systems (GPFS) used for supercomputers and increasingly big analytics jobs.
One of the questions that can be asked is how important is the small to midsize business market to IBM? Certainly the SMB is home to at least 80 percent and maybe even 90 percent of IBM i shops. So it is important in that respect. But does getting out of the X86 server business hurt IBM’s penetration of that market and potentially hurt IBM i?
“I do not believe that it will negatively affect the IBM Power i business,” says Jim Kandrac, the founder and president of United Computer Group, an IBM business partner located in Independence, Ohio. “Power and Intel really haven’t played together in years. In some cases, UCG provided the ‘X’ as a service to clients. In other cases, the clients sourced their own, because of a brand preference, and it was handled by a different IT unit. Many mid-market shops have different sandboxes and i and Intel boxes are separate. Due to politics or religion, those purchases are separate in most cases. I view IBM selling off X as a proper strategic move. It apparently was not ‘core’ to the growth of IBM and at the end of the day it may be better for IBM, their clients, Lenovo, and partners. Unless hardware is blended with software and services, it is a commodity.”
Apparently IBM was unable to make that blend, because X86 is definitely a commodity. Eating sandwiches that are mostly bread and light on the meat is not an appetizing diet for Big Blue, who nonetheless is looking to lose some weight. Approximately 7,500 employees will be trimmed from the IBM payroll when this deal is finalized. So Lenovo gets the task of finding more meat for the X86 sandwich and IBM will fill up on software and services, if it can make that work. To Lenovo’s credit, it found more meat in IBM’s PC business. If they do the same with the X86 server business, and IBM picks up the software and services numbers it needs, then this looks pretty smart.
It could backfire though if IBM loses the somewhat tenuous grip it has in the SMB market.
“I think IBM’s sale of the Intel business to Lenovo is a long term disaster for IBM and Power in the SMB,” says Doug Fulmer, an ex-IBMer with strong ties to the IBM i on Power. He works exclusively with SMB companies as a representative of IBM business partner KS2 Technologies located in Grapevine, Texas. “I believe that IBM is only successful to the extent that they penetrate a business. Virtually every business is at least looking at Windows and Intel, and if IBM has no presence in the Intel space, they lose their presence and penetration, and then will lose hardware, software and services business, as that penetration is usurped by Dell and HP in the United States and other brands abroad.”
“Intel was a loss-leader for IBM, but I think they are too arrogant to realize it,” Fulmer continues. “Without the foothold that their hardware business provides, their software and services will wither. IBM is too expensive and too hard to do business with. I think that IBM thinks they can beat Microsoft and Intel with Linux on Power. And I think that they are bonkers. Other than one customer in west Texas, I haven’t seen anyone in SMB using Linux for anything in my neck of the woods. It’s all Windows.”
Then there’s the concern that if IBM doesn’t lead with hardware in small and midsize companies, will its software and services be appropriately tailored for this price-sensitive and simple solutions demanding market?
“I hope IBM is going to refocus its efforts into designing bundled midmarket solutions and put more resources and priorities into the midmarket,” says Pete Elliot, director of marketing for Key Information Systems, an IBM business partner. “They need to do something in the midmarket. That is where the growth is coming. IBM gets most of its income from the high end–the large enterprise users. It needs to grow in the SMB. There needs to be more entrance points to the midmarket with Power Systems, software, and solutions. There needs to be integration and a long-term game plan.”
Key has been successful by working with shops that are interested in the X86 architecture. It bundles software solutions and some services around X86 and virtualization. It is a reasonable entry point for companies to get into the IBM family, Elliot says. Approximately 25 percent of the revenue at Key comes from IBM products used by midsize businesses.
Elliot gives a lot of credit for Key’s past success in the SMB to Avnet, the company that oversees the distribution channel that Key belongs to.
“Avnet has done a great job of supporting the X86,” Elliot says. “They are fully onboard with Lenovo–not just because they have to be, but because the level of support will be as good and possibly better.”
It’s true Lenovo might be better at this than IBM. It might be more focused. The SMB might be a bigger priority. And the sales channel is impressive, although it is adjusting to a new way of doing business that doesn’t depend on hardware sales like it used to.
“Avnet knows what needs to be done in order to assure continuity in the process,” Elliot says. “I’m an optimist. I gotta believe that when we go to a customer with a Lenovo X86 server instead of an IBM X86 server that the value proposition is still very strong.”
Tony Madden, senior vice president and supplier business executive for Avnet Technology Solutions was out of the country on business last week, but responded via email to my question about the X86 business transition.
“It’s too early to tell how the recently announced plans for Lenovo to acquire IBM’s x86 server business will affect the IBM Power Systems market. At this time, it appears that the pending deal will re-emphasize IBM’s focus on Power Systems, which we believe would be beneficial for our channel partners in this market.”
Madden also noted that IBM continues to focus on higher value-oriented solutions, which aligns with Avnet’s strategy as a solutions distributor and provides competitive advantage.
At the IBM-Avnet business partner event in August, one of the most emphasized business solutions was converged systems. IBM calls its version of converged systems PureFlex. It is designed to integrate multiple environments–IBM i on Power and Windows on X86 are the obvious ones for IBM midrange shops–inside a common hardware infrastructure that includes compute, storage, networking, virtualization, and systems management. It takes what BladeCenter started, improves upon it, and broadens the scope to include all infrastructure, with the cloud being part of the picture. The main point of this is to relieve companies from the infrastructure building process, which allows them to focus on applications, which can be run on the platform that makes the most sense be that IBM i, AIX, Linux, or Windows. And instead of building infrastructure for each of those platforms, it’s set up once and serves all platforms.
X86 plays a big part in this and with IBM stepping out of the X86 server business, there is concern that the converged systems program will face a setback. Even though it had yet to build momentum in the IBM i community, there is hope that it could settle into an attractive price point while providing ease of management for both systems.
“A year ago, IBM was pounding on partners like us to get skilled up on Pure,” Fulmer says. “Now they sell off Intel, which was the foundation of Pure. IBM is making noises about OEM-ing Lenovo Intel into Pure, but they will have lost the mind share of customers accustomed to the IBM brand and having a single throat to choke.”
Fulmer also insists PureFlex is too expensive and too complex for the SMB market and companies prefer the standard rack and stack arrangement. “Pure as a Power-only play is a non-starter,” he says. “Unless a customer was going to integrate Intel and Power in a common rack and storage, Pure is a much less compelling argument.”
It is apparent that X86 will continue to be important in small to midsize companies. By taking itself out of the game IBM has put a great deal of trust in Lenovo and the channel to maintain this important portal into the SMB market.
“I view IBM selling off X as a proper strategic move,” Kandrac says. “It apparently was not ‘core’ to the growth of IBM, and at the end of the day it may be better for IBM, its clients, Lenovo, and the business partners.”