Intel has just filed a lawsuit against NVIDIA, claiming that the front side bus license that NVIDIA is using to build Intel-compatible chipset for pre-Core i7 CPUs is not valid on Core i7 and future CPUs. NVIDIA Just responded:
NVIDIA believes that our bus license with Intel clearly enables us to build chipsets for Intel CPUs with integrated memory controllers. We are aggressively developing new products for Intel’s current front side bus (MCP79 and MCP89) and for Intel’s future bus, DMI.
The soul of the PC has become the GPU and the CPU is becoming less relevant. Intel is trying to prevent GPU adoption since the evidence is undeniable that the CPU has run its course. The rapid shift to the smallest and lowest price CPUs like Atom is a clear reflection of this trend. (Read the full press release)
We’re not familiar enough with the details of the license, but just to give you some context: one might think that a few years ago, NVIDIA had build so much pressure on Intel with its AMD-only multi-GPU SLI that Intel “had to” give NVIDIA a front-side bus (the connection between the CPU and the rest of the motherboard) license so that NVIDIA could produce SLI-enabled chipsets for Intel CPUs.
Since then, NVIDIA has become not only a dominant graphics company, but also a rapidly visible “compute” company, thanks to CUDA, its general-purpose computing platform. It is generally believed that Intel sees GPU-compute a possible threat, not to the “existence” of the CPU, but to its “relevance” – this is a distinction that you need to understand well. The CPU isn’t going away, but if it is less relevant, that means that it will be hard to charge a premium for it.
Intel will come up with Larrabee, a stand-alone GPU that should also be a formidable compute platform, at least on paper. No working silicon has been demonstrated yet.
Back to today’s event: It could be that Intel simply wants to enforce its licensing terms. It’s business after all. It could also be that Intel wants to do everything it can to prevent NVIDIA from building chipsets that would contain an NVIDIA GPU that can process things that would otherwise be processed by the CPU. Finally, remember that Intel can now build SLI-enabled motherboards (at a price), so it appears that it doesn’t need NVIDIA that much. AMD is also in no position to be a threat on the CPU side, so NVIDIA can’t build much pressure like it did a few years back. Don’t forget that AMD since bought NVIDIA’s arch-rival ATI…
As we said above, we’re not lawyers, and we’re not familiar with the details of the license. Who’s right or wrong will be decided by the judge. In the court of public opinion, we guess that this is bad PR for Intel, but business is business. What we do know is that NVIDIA and Intel are clearly two “compute” companies that will likely be friends and foes, depending on which division of the companies you talk to, for years to come. Just read what NVIDIA’s CEO (Jen-Hsun Huang) has to say:
“We are confident that our license, as negotiated, applies,” “At the heart of this issue is that the CPU has run its course and the soul of the PC is shifting quickly to the GPU. This is clearly an attempt to stifle innovation to protect a decaying CPU business.”
Update: We’ve learned that Intel has been discussing this matter with NVIDIA for about a year. As both companies could not agree, Intel is asking the court to decide.