Intel will gain access to Nvidia's patents while paying the graphics chip supplier $1.5 billion in licensing fees as part of a new six-year agreement.
"For the future use of Nvidia's technology, Intel will pay Nvidia an aggregate of $1.5 billion in licensing fees payable in five annual installments, beginning Jan. 18, 2011," Nvidia announced today.
Furthermore, Nvidia and Intel have agreed to drop all outstanding legal disputes between them.
The crux of the agreement is that Intel gains access to all of Nvidia's GPU (graphics processing unit) patents but Nvidia gains access to only certain Intel patents. To compensate for the lop-sided patent access (which favors Intel), Intel pays Nvidia $1.5 billion.
Intel and Nvidia had both sued each other in early 2009 in a dispute that originally centered on a chipset license agreement. Intel had contended the cross license does not extend to Intel's future-generation processors, and Nvidia countersued blocking access to its patent portfolio.
In effect, Nvidia was barred from building Intel-compatible chipsets beyond the Core 2 Duo generation of processors. For example, the second generation of Apple's MacBook Air used an Nvidia chipset along with Intel's Core 2 Duo processor. However, Nvidia could not build chipsets for the newest generation of Intel Core i3, i5, and i7 processors. This, in effect, forced Apple to stay with Intel's older-generation Core 2 Duo processors in its newest MacBook Airs because it allowed Apple to legally continue to use Nvidia chipsets.
The agreement announced Monday still bars Nvidia from using any of Intel's x86 technology and, as a result, Nvidia cannot build x86-compatible chipsets, according to Intel. But Nvidia CEO Jen-Hsun Huang made it clear he's not interested. "We've already said many times that we have no intention to build chipsets for Intel processors," he said in the conference call Monday afternoon. And many PC makers (including Apple) still use discrete (standalone) Nvidia graphics processing units (GPUs) that attach to Intel chipsets.
Huang expounded on its traditional strong suit, GPUs--which hold patents that Intel is paying for and which Nvidia incorporates into its ARM processors. GPUs excel at parallel processing, whereas CPUs (central processing units)--such as Intel's x86 chips--do sequential processing. Both types of processors have their merits, though GPUs have the potential to be much faster than CPUs at doing visual processing and scientific number-crunching, for example.
"I don't think you can build a modern computer today without a state-of-the-art GPU technology. Anytime you can do something in parallel, it's better than sequential," Huang said.
"Our cross license with Intel reflects the substantial value of our visual and parallel computing technologies. It also underscores the importance of our inventions to the future of personal computing, as well as the expanding markets for mobile and cloud computing," said Huang in an official statement today.
Huang went on to say the company's focus is now on ARM processors--which compete with Intel's x86 chips in small devices like Netbooks and tablets. "It's a foregone conclusion that ARM is the most important [chip] architecture. ARM will be the largest installed-based processor. It's pervasive and open. We will extend the ARM processor with our GPU," he said.
Huang pointed to Microsoft's announcement at CES to port its next major release of Windows to ARM processors and Nvidia's CES announcement of Project Denver, in which it will design high-performance ARM chips for desktops and supercomputers. Those future Nvidia chips will be hybrids--much like Intel's just-announced Sandy Bridge processor. "Project Denver...features an Nvidia CPU running the ARM instruction set, which will be fully integrated on the same chip as the Nvidia GPU," Bill Dally, Nvidia's chief scientist, said last week.
Updated at 3:25 p.m. PST throughout.