The PC will die
I expect one thing to happen that the industry will have no defence against: The PC is going to be, for the most part replaced. You may think this a somewhat bold assertion given it's success, but in order to understand why you must first understand that the economics of the PC market are somewhat different compared to the rest of the world, this alone almost makes the PC's replacement a forgone conclusion.
When PC and Mac fans duke it out on-line the Mac is commonly criticised as being propriety hardware whereas the PC is made up of cheap commodity hardware. This argument hides a truth rarely revealed: The PC itself is a highly expensive niche product. x86 processors power more than 90% of the desktop market and have a sizeable chunk of the server market. Yet these same processors make up just 3% of the total number of CPUs shipped per year.
The CPU industry as a whole ships more embedded CPUs per year than the number of people on the planet - and they have done this for the last 8 years running. Ubiquitous computing is not coming, it is here already, but no one seems to have noticed. Phones, Cameras, Hi-Fis, TVs, DVD players, MP3 players - they all have CPUs. These use 32 bit processors which are only a minority in the CPU market, the real action  is 8 bit CPUs which are the market leader , this is liable to inspire most computer enthusiasts with horror but just consider that the embedded industry spent the last decade upgrading from 4 bit CPUs!
Despite it's relatively tiny size the PC industry gets the majority of the profits due to some very large margins. What's more, Microsoft have margins which dwarf the computer or semiconductor manufacturers, including 228% in their "InfoWorker" (read Office) division and a whopping 415% profit on Windows, by contrast Dell made 6% (figures for 3rd Quater 2003) .
It is these huge margins combined with the under-utilised capabilities of modern microprocessors that mean the x86 and especially Microsoft are vulnerable. It would be very easy to get into a price war with the PC industry - and win.
A second difference from the rest of the world is that PCs need to constantly upgraded. The entire industry is based on obsolescence. With a constant flow of people and businesses replacing computers with new ones Computer, Semiconductor, Operating System and Application vendors have done very nicely thank you over the past 20 years. However computers and software have now become so powerful and feature rich that the endless upgrade cycle is beginning to come to an end. In other industries users only upgrade when the product becomes useless, TVs and other home appliances can be in use for many years before needing to be replaced. This is also starting to become true for the Computer, I have a 800MHz PC and it's just fine, I could upgrade but the fact is I have no pressing need to.
The challenge that the PC is going to have to face is a computer not based on high performance, high margins or built in obsolescence, it will be a small, ultra-low cost computer.
Return of the Mac
The original idea for the Macintosh was for an ultra low cost computer, it would have been very different had that vision of been followed, It would of had an 8 bit CPU, no mouse and no GUI. The end product was changed drastically after Steve Jobs  took over as the project leader and he changed it into the vastly better product later launched.
Ultra low cost computers have been done before, the PC Jr  was one example. By today's standards it's $1000+ price tag is not exactly low cost but then it was something of a bargain. However users then needed all the power they could get and it subsequently never sold.
A low cost computer is unlikely to be of interest to most computer enthusiasts but that is irrelevant since they only buy a fraction of the computers sold. Most computers go into offices where they do exciting tasks like e-mail, web browsing, word processing, presentations and spreadsheets. You do not need a 1GHz CPU to do these tasks, in fact most of these tasks have been done on much, much slower computers for many years, a 100MHz processor will do fine, but our low cost system wont be quite that slow.
Someday someone with enough resources is going to start building them, I expect them to arrive from Asian countries most likely running a form of Linux. Education will see the immediate benefit followed by governments, business users will follow when they see the millions saved in large rollouts. Slowly at first then faster they will start eating away at the PC market, they will eventually take over from below. Microsoft will no doubt fight back with bigger, better software with ever more features but is this really important in business? Microsoft will have to have products good enough to justify their colossal margins, with low cost competition they are going to have to fight hard to avoid having to slash prices. The strategy I suggested they are contemplating for PCs in Part 2 will not help them here, they will not be able to get special hardware added to these boxes.