If Intel is about to jump back into the graphics market again, where does that leave Nvidia and is it time it joined the AMD/Intel twosome?

It seems to have been coming for the last year or so but it is clear that Intel’s Larrabee project is slowing coming to fruition – something that probably won’t have Nvidia executives doing handstands.

The gaming GPU manufacturer has had a pretty good run of late with the release of its new GeForce GTX200-series GPUs but the big money market for graphics chip vendors has traditionally been the OEM (original equipment manufacture) deals it does with big-name notebook vendors such as Toshiba and HP.

These deals have long been the bread-and-butter for GPU manufacturers but with Intel seemingly about to create graphics solutions of its own that may just rival Nvidia’s, you have to wonder just what the future holds for Nvidia.

The buyout of ATI ensured AMD could now supply complete system solutions to big-name vendors, something those vendors are typically keen to have. They aren’t really interested in having to source their parts from too many vendors so a “one-stop shop” for CPU/chipset/graphics makes sense and to some extent, it appears to be working for AMD.

Up until now, Intel’s Express chipset graphics solutions have been, in a word, woeful – at least in terms of 3D gaming and accelerated video playback. However, Larrabee is being seen as a major step forward for Intel that might bring them back into the gaming market as a viable player.

“Might” however is the operative word here.

However, sticking with a macro look at things for the moment, say Intel actually did hit upon a winner with Larrabee. Where does that leave Nvidia?

While it will still have its market brand and the ability to churn out high-performance parts, it could likely lose its lucrative OEM market, particularly if Intel prices Larrabee aggressively, something the chip giant hasn’t been afraid of doing in the past.

Nvidia already produces graphics and chipset solutions and with its ability to develop complex GPUs, moving that technology to CPU wouldn’t be out of the realms of possibility.

Of course, the idea of an Nvidia CPU isn’t new and was reported as far back as 2006 when the GPU manufacturer bought a small start-up called Stexar. Interestingly, it was about that time early reports suggested a possible 2008 release for such a CPU.

On May 26 this year, Nvidia announced the release of the Tegra APX 2500 CPU, an ARM-based chip aimed at pretty much the same market – PDAs, smartphones etc. And although it has some analysts thinking bigger things are ahead for Nvidia in the CPU market, the major problem so far has been Nvidia’s apparent lack of a license to create x86 architecture-based processors, which effectively rules out use of Windows. And without support for Windows, you really don’t have a desktop CPU.

Rumours have also popped up that Nvidia could purchase chipset and CPU competitor VIA Technologies and gain access to x86 licenses that way. So far, it’s been “non-denial denials” from everyone involved. The problem here however is that VIA really hasn’t pumped out a decent x86 chip for some time, let alone any serious multi-core CPUs so it would be hard to see a VIA purchase as anything more than “one step backward to go two steps forward” at best.

The latest rumours eminating this week are that Nvidia is indeed working on an x86 processor, although you won’t get anyone from Nvidia to confirm this. The real question is if this is the case, whether or not it has the x86 licenses to go with it.

If not, expect AMD and Intel to go in hard with law suits and all guns blazing.

The story behind Stexar, apparently full of ex-Intel techs, is an interesting one and will have given Nvidia the know-how but unlikely the necessary license.

If all this is true, it makes an Nvidia move into the x86 CPU market for Nvidia as much a gamble as it would be a calculated risk.

But given the rumours won’t go away, one thing is for certain – the next three months should be interesting.

---