In a move that’s sent ripples through the tech world, Nvidia has just announced it’s ploughing a massive USD5 billion (~RM21 billion) into Intel. That massive investment nets them roughly a 4% stake in the chipmaking titan, making Nvidia one of Intel’s largest shareholders.
But it’s not just about the money. The two companies have also inked a new pact to jointly develop chips for both personal computers and data centres. Interestingly, the deal specifically leaves out Intel’s contract manufacturing, or ‘foundry’, business. So, don’t expect Intel to start fabricating chips for Nvidia anytime soon.
While they’re keeping the precise financial terms under wraps, the pair have committed to creating ‘multiple generations’ of future products together. Officials from both camps were keen to frame this as a straight-up commercial arrangement – a case of swapping chips to build better products, with no licensing element involved.
So, what’s actually being built? For the data centre, Intel will design custom central processors (CPUs) that Nvidia plans to bundle with its own AI accelerators, through Nvidia’s GPUs. To make sure these chips work together seamlessly, they’ll use Nvidia’s proprietary NVLink technology to communicate at seriously high speeds.
On the consumer side, the favour is returned. Nvidia will provide Intel with a custom graphics chip to package alongside its PC CPUs, also linked up with that same speedy connection. This could be a real game-changer, potentially giving Intel a leg up against competitors like AMD in the never-ending processor wars.