Nvidia not only develops hardware, they also put a lot of focus on software and platforms for the hardware they develop. The software portfolio includes CUDA, an abbreviation for Compute Unified Device Architecture. CUDA is a platform and framework for building applications to take advantage of the computing power of graphics hardware, in this case only in combination with Nvidia’s products.
Letting other players in by making the CUDA platform open source is not on the company’s agenda. Jensen Huang, Nvidia’s CEO, explains to The Register that they are open to others building hardware compatible with CUDA. Huang explains that this is not a new approach, but no one has ever asked the question – at least none of significant size.
Underneath CUDA is Nvidia’s hardware. There’s really nothing to open source. If somebody would like to build an application for CUDA or to build another chip for CUDA, we’re not not fundamentally against it, and nobody has ever asked. – Jensen Huang
Nvidia also does not intend to cast its own hardware open source for competitors to take part in, and those who want to use CUDA are forced to buy Nvidia’s products at present. Google and Meta (formerly Facebook) are examples of players who have the muscle to develop their own circles for different purposes, especially if the offer on the market does not fit. The Register writes that it is customers of that caliber who can get Nvidia to invite them to create CUDA-compatible alternatives.
Such a move would mean that the electronics giants could continue to benefit from Nvidia’s well-polished and well-established framework, but with circuits that fit the application like a glove. It can also be assumed that it is necessary for Nvidia to meet such large customers, as the alternative is that the resources are instead spent on alternative solutions and to completely abandon Nvidia’s software and hardware. Whether some of the industry’s big dragons work on such graphics circuits, however, remains to be seen.