The Inquirer claims that Nvidia is preparing to produce it's own CPU. I don't know. I've never heard of this Stexar thing before.

Anyway, they say the following about the problems with non-x86 compatible CPUs, and they're mostly right:

To step aside for a moment, let's ask this question. If Nvidia does not make an X86 CPU, but an ARM, PPC or some other ISA? Well, it is, again more than capable, but that would mean giving up the market it has a pretty solid lock on now. Buy the new Nvidia CPU, it doesn't run Windows and won't run your software or games, but the theoretical graphics power is astounding! I can see the lines forming to buy one as soon as they are released. Not. Linux is very cool, but basing a mainstream CPU only on it may not be a bright business move.

OK. But x86 is an ancient, ugly kludge. The chips are horrifically inefficient because they have to support instructions no one has used in 20 years. Plus there's the BIOS, the design of which is nearly that old by itself, and which is only now beginning to be replaced.

I’ve finally learned what “upward compatible” means. It means we get to keep all our old mistakes.
– Dennie van Tassel

Then there's Windows. It suffers from this architecture, but worse, it adds it's own layers of obsolete APIs and bugs intentionally preserved and screensavers that ship in original Windows 3.1 form with XP. Vista won't fix it. Vista will make it worse.

It doesn't have to be this way, though. Check out the list of architectures supported by a decent Linux distro, such as Debian. All the software that ships with these systems works on all the architectures supported. Hardware compatibility isn't necessary, because all the source code is available and portable.

So, how can we fix this? How can we free the computer industry from the tar pit it's mired in?

Emulation. There are already a number of low-power processors that emulate x86 entirely in software. Suppose an enterprising CPU manufacturer created an extremely high performance CPU, with a huge number of cores. Suppose they built a software x86 emulation layer that allowed the OS to run natively while some threads ran emulated x86 code. You could even imagine that the released this software under the GNU GPL and integrated it into the Linux kernel, so that it became platform independent.

The next step is obvious: WINE. A reimplementation of the many Windows APIs that runs on Unix.

Now we're set. We can run our legacy apps, DirectX games included, on any platform at all, so long as it has a Linux port. Our hypothetical manufacturer starts churning out massively parallel, high performance legacy-free boxes for a third the cost of a conventional x86 machine. They sell, because people can use them to do what they want. A competing manufacturer can ship an entirely different hardware platform, this one optimized at the instruction-set level for video processing, and people still buy it because they can run all the same software on this one as on the other.