Linus Torvalds Called Nvidia "The Single Worst Company We've Ever Dealt With"

Full text

For as long as there's been software there has been tension between open source and closed source. The creator of the Linux kernel, software engineer Linus Torvalds, has long been an outspoken advocate for open source environments – while putting the boot into closed source corporate entities who have tried to exploit the ecosystem.

Open for business

During a Q&A at Aalto University, an attendee complained that the Nvidia Optimus chip – a processor that switches between integrated graphics and dedicated graphics – was no longer being supported on Linux.

Quote of the day

This article is part of TechRadar Pro's QOTD project to provide an insight into the minds of the brightest and most recognized figures in the technology industry today and in years gone by. Read the full series here .

That meant that this expensive chip fitted into her Linux laptop was basically useless unless developers worked on a fix by reverse-engineering the process. Torvalds answered by giving Nvidia both barrels – complaining that the company was profiting from the Linux Foundation ecosystem through products like the Tegra chips for Android devices.

At the time, he complained that competitors like Intel and AMD were happy to co-operate with Linux to build native drivers whereas Nvidia would treat its proprietary code as classified material. This exchange culminated with an angry Torvalds flipping Nvidia off to camera.

Turning the tables

The situation today is almost entirely different from that we found nearly 15 years ago. Nvidia has embraced open source with open arms, and has also fully transitioned to open source Linux GPU kernel modules as being the default options as of 2022, pushed under a dual GPL/MIT licence .

With the AI buildout also fully underway, Nvidia has crafted the Linux-based DGX OS for its personal mini supercomputing DGX Spark products. This Ubuntu-based system is designed to work with the entire Nvidia software stack – with this AI layer (including software like CUDA, cuDNN and NCCL) being closed.

Although the battle between open and closed source remains active, Nvidia's reversal signalled that the future of enterprise and AI-centric computing was to be found in open source environments.