Don't forget also that:

Mainstream use of Linux, probably the most important software that takes full advantage of advanced features, increased demand for dual core.

Video editing on the desktop also is far more common.

Intel is very well positioned to compete with AMD, since it has closed the price/performance gap, no mean feat while you're still doing R&D.

The shift to powered Ethernet and energy efficient Ethernet (which sends shorter bursts of traffic at higher speeds and then goes dormant for longer) mean a lot more network chip salesare poorly coordinated and also poorly coordinated with including devices (security cams, door locks) that never had them before. The engineering of these is difficult especially since the standards are poorly coordinated and even the 40Gb and 100Gb Ethernet groups are still fighting. It takes an 800 lb gorilla like Intel to cut this Gordian knot, especially as no one is following Cisco's lead any more. For good reason: their PoE implementations have been known to go on by surprise and fry hosting racks which by definition operate at the maximum heat they can stand already. As in the early days of Ethernet it will be safest to make sure you buy all your chips and switches from one vendor. But who makes both chips and switches any more? Hm.

Also the motherboard business is going to pick up for various reasons including the cull of vendors who can't handle the demands of the very sensitive higher speed chips and the tricky problems of power supply to many USB and PoE ports and handling hundreds of megabits of BPL data on the AC power plug when IEEE P1901 is done. Plus the Climate Savers and Energy Star configurations that flatline the board but still expect it to hear someone whispering one bit at it. Good luck AMD/ATI teaching motherboard makers how to deal with all this, with no motherboard team inhouse. Some problems just require daily contact between people working for the same company, and motherboards that feed a lot of power to a lot of peripherals and have to deal with gigabit data feeds on multiple buses that put power on the same wires may be one of those. So is any optical chip on the motherboard. Or any use of quantum computing devices beyond one chip.

Also as the PC architecture finally taps out, as it causes more problems like the 40Gb vs. 100Gb Ethernet fight (40Gb is the maximum PCI-E can do), Intel can play its network and motherboard talent to set the post-PC standard for LAN-on-motherboard. Now that Ethernet handles DC power up to 80 watts per device under 802.3at, there's no good reason not to implement a backplane that is essentially just a 100Gb Ethernet. Maybe even 200Gb to 800Gb using parallel Ethernets which the memory, GPU and CPU sit on directly via a blade. While an outward-facing blade has the various bus controllers, the power supply or at least its brain, and the terabit fiber optic storage and network interface, plus probably the wireless module and simultaneous decoding of thousands of HDTV channels to find something worth watching - this alone will require a little OS like VMware on a dinky chip on that blade, which can also do the firewalling and filtering and authentication. Intel can make both these blades now. AMD can only make the inner one so far.

The x86 architecture is a problem when trying to match Sun's energy performance or beat Cisco at NOC application. But as GPUs and dedicated I/O processors take over more of the work, this matters less and less. Emulation is good enough now for Intel to just abandon x86 hardware any time. They have more engineers than AMD does so they can make it hard for AMD to follow where they lead, especially if they set the bus standard or win the home theater wars with ViiV.

So yeah be bullish on Intel but watch them closely on mobile phones and laptops - how much is Intel getting paid for each one of those? They can rule the high end but can lose the low end if Celeron doesn't start showing up on phones. Apple took too long to become a phone company. Intel might too.

Finally consider that each office or home might be done with upgrading desktops. Your next PC is probably your last. Time for Intel to look ahead at what Sony, Samsung, LG, Google, Apple, Lenovo, Nintendo, etc. are doing, which is about small devices held in your hand or hung on your belt, headphones, and big screens on some one else's wall. Using whatever CPU and GPU and bandwidth someone else is not using right now. A computer is no longer a box. It's a very temporary lease on a few devices that will all have IPv6 numbers.