Nvidia appears to be radically rethinking its Tegra business model. Writing in his bog, company spokesman David Shannon said Nvidia’s “next step” is to license GPU cores and visual computing patents to other outfits.
“We'll start by licensing the GPU core based on the Nvidia Kepler architecture, the world's most advanced, most efficient GPU. Its DX11, OpenGL 4.3, and GPGPU capabilities, along with vastly superior performance and efficiency, create a new class of licensable GPU cores,” he said.
“Through our efforts designing Tegra into mobile devices, we've gained valuable experience designing for the smallest power envelopes. As a result, Kepler can operate in a half-watt power envelope, making it scalable from smartphones to supercomputers.”
It is a rather surprising change of heart on Nvidia’s part. Although the GPUs used in the Tegra 2 and Tegra 3 weren’t very competitive, the new graphics subsystem in the Tegra 4 and Tegra 4i seems to be a lot more capable than anything the competition has to offer. Next generation Kepler-based cores should be even better.
At the moment the vast majority of ARM SoCs ships with Imagination Technologies or ARM GPUs and Nvidia doesn’t have much of an overall market share. However, this could change in a heartbeat and Nvidia believes it could cash in. On the other hand, Nvidia’s trump card for next generation Tegra designs might not be Nvidia exclusive anymore, as other outfits could get the technology.
No Tegra chip has flopped so far, but they weren't a huge success, either. Scoring a GPU IP deal for Samsung’s Exynos SoCs or Apple’s A-series chips might be more lucrative for Nvidia in the long run, even in such an approach undermines Tegra’s competitiveness.