With the pervasive rise of smartphones, tablets, and high-def screens, the market for visual computing has grown beyond what NVIDIA (NVDA 3.65%) could realistically address with its own product lines alone. To keep up with the rapid pace of growth, the company has decided to license its cutting-edge Kepler GPU architecture to device manufacturers.

Not only does this move give device manufacturers an opportunity to leverage NVIDIA's GPU technology, it also gives NVIDIA an opportunity to expand its addressable market and create new revenue streams. In an age where over a billion computing devices are produced yearly, NVIDA's decision seems to be right on the money.

In fact, it almost sounds too good to be true.

A conflicted place
As a result of this development, NVIDIA's business model has transformed somewhere in between ARM Holdings' and Intel's core business model, which isn't necessarily a good thing. Because NVIDIA will be licensing and selling GPU products, it could create a breeding ground for conflicts of interest between potential customers and the company's own interests. What's going to stop NVIDIA from developing a markedly better GPU that it hoards for itself if the company thinks it could make more money that way versus licensing the technology out?

Since ARM only licenses out its technology and doesn't compete with its customers, it gives its customers the peace of mind that they will get ARM's best technology.

Although Intel licenses out its x86 architecture to a handful of semiconductor companies, the company has zero conflicts of interest as far as device manufacturers are concerned. Considering the vast majority of Intel's business relies on relationships with device manufacturers, it's wise to keep any potential conflicts out of the equation.

It's never this simple
Although NVIDIA's announcement opens up a world of opportunities for the company to land high-profile clients like Apple, Samsung, or possibly even Intel, I wouldn't be betting on the chances of it happening anytime soon. It's likely there will be a host of engineering challenges associated with manufacturing NVIDIA's technology on a different foundry process. In other words, if NVIDIA's potential customer isn't already a Taiwan Semiconductor customer like NVIDIA, it may not be worthwhile for that customer to license NVIDIA's tech. Since Apple, Samsung, and Intel aren't currently TSM customers, it likely diminishes the possibility of such a deal being struck with NVIDIA anytime soon.

No matter how great an idea looks on paper, there's always a chance the real world plays out differently.