First the PCs, now System X: Is IBM ready to flog server biz to Lenovo?
Speculation mounts over elective X86-ectomy...
HPC blog Is IBM selling all or part of its x86 server biz to PC giant Lenovo? Nothing has been confirmed by either party, but it sure looks like there’s something in the works. So what might this deal look like?
I don’t see IBM selling off the entire System x division to Lenovo. What they’ll sell is the most commoditised and least differentiated boxes in their product line. This means goodbye to their single- and dual-socket tower servers for sure. It also means that they’ll probably bid farewell to almost all of the models in their M3 and M4 rackmount server lineup. Selling off this part of System x removes sixteen SKUs. So what will IBM keep?
I think they’ll hold onto their blade business for sure. There are good margins there, and enough differentiation to stand out from competitors. They should also hold onto their iDataPlex product group. iDataPlex is a very dense system that was designed from the ground up to appeal to the scale-out crowd who buy their servers in 10,000-system lots. It has also made a name for itself in HPC, with 30 of the systems on the Top 500 list.
I think that IBM’s x5 systems will also stay with Big Blue. These are the systems that use IBM’s custom x86 chipset (dubbed X-Architecture), which gives them the ability to add features and functions that are unique to IBM. While I don’t think these systems are IBM’s biggest volume sellers (IBM doesn’t disclose this info), they do carry margins that are significantly higher than the other System x boxes.
It’s doubtful that IBM would sell off any of the System x components in its newly introduced Flex and PureFlex initiative. These are highly integrated packages of server, network, storage, o/s, virtualisation and management software sold as a complete package. In the x86 version of Flex/PureFlex, customers can choose from three nodes, two dual-socket choices, and one quad-socket alternative. The main difference between the two dual-processor systems is the number of DIMM slots (12 vs. 24).
So if IBM sells off what I think it’ll sell off, it’ll be shedding around 16 mostly low-end systems and transferring the product IP, people, and production facilities to Lenovo. Post-sale IBM x86 products would include five rackmount servers (x5 models plus Flex models), four blade chassis, and four blade flavors, plus iDataPlex. I would also expect that any deal would include IBM reselling (and rebadging) low-end Lenovo servers to IBM legacy customers.
The deal as described above keeps IBM in the x86 game enough to continue innovation and skim the cream from the high end of the market. IBM’s average margin on x86 systems will rise significantly after they jettison the low-end and less specialised gear. They’ll also get a cash infusion when Lenovo’s multi-billion dollar payment hits their checking account. One potential downside, however, is that this might make it more difficult for IBM to compete in some large commodity-gear oriented markets like HPC.
This deal gives Lenovo the chance to duplicate what they did with IBM’s PC division – shine it up and turn it into a money maker. With PCs (particularly Thinkpads), I think Lenovo has managed to keep or enhance most of the IBM-like qualities that appealed to users while dropping prices to the point where they are competitive with other name-brand systems.
Lenovo has a better-than-decent shot to do the same thing with larger systems. The Chinese home market has an insatiable appetite for servers, and Lenovo definitely knows how to sell in China (and greater Asia too). If this deal happens, Lenovo will have a much broader line of systems to sell, all with the IBM engineering heritage behind them – but at Lenovo prices. That, along with continuing volume from IBM, could be enough to build Lenovo into at least a regional server power.