It is a lot easier for IBM, EMC and Cisco to reduce the number of needed reboots. They own the hardware and the software so they’ve got total control. With client and server OSs (Windows, *nix, etc) the hardware is built by several companies and the software is built by one or more companies. Because of this coupling there are sadly going to be bugs which may require reboots to fix.
The OS vendors have gotten a lot better about the need for these reboots, with the exception being when patches are applied which need to update files which are in use. It is common for Windows servers to run for months without a reboot (unless a patch is deployed). *nix servers can easily stay online for months or years between reboots.