Making a smartphone used to be both a logistical and an engineering nightmare. You had to buy multiple components from multiple vendors and then hire an army of engineers to write the machine code needed to make all those chips talk to each other. But then Qualcomm, in 2008, introduced Snapdragon. Their sales pitch was great: We’ll give you, the phone maker, everything you need to make a device. All you need to do is throw in a screen, battery, and shove it inside a plastic case. And since then, the company has been untouchable.

Enter NVIDIA, who saw Qualcomm printing money hand over fist and decided they want in on the action. They launched the Tegra brand with the hope that the combination of ARM’s powerful CPUs and their in-house developed GPUs would set the world on fire. That didn’t happen. So NVIDIA bought a company called Icera in order to integrate cellular modems inside Tegra. Did that help? No, not really.

According to Reuters, the Tegra business is floundering, to the point where sales are down 54% year over year. Analysts from Evercore and Needham are urging NVIDIA to “spin out Tegra or sell the business related to phones and tablets”, because the mobile division is hurting the rest of the company.

Should NVIDIA quit? I’m going to have to say yes. Qualcomm is already getting enough competition from Chinese players such as MediaTek and Rockchip. More importantly, Intel is hungry, and there is little doubt in my mind that they’ll have a solution that’ll be competitive with Snapdragon out on the market at some point in the future. Maybe not this year, or the year after that, but eventually they will, and that’ll come to bite Qualcomm in the ass.

So yes, maybe NVIDIA should listen and go back to focusing on GPUs.