While Intel has been experiencing rough times of late, there might be some good news on the horizon for its foundry business. Rather than just making chips for its own product and a few other companies, a report is claiming that Intel is currently working with Broadcom and Nvidia to see if its 18A process node is good enough for high-end GPUs, SoCs, and other critical processors.
Don't put your eggs in one basket is probably the most apt idiom for how chip giants are approaching manufacturing at the moment. That's because when it comes to CPUs, GPUs, APUs, and SoCs for desktop PCs, laptops, tablets, phones, and consoles, the vast majority of such processors are all made by one company: TSMC.
As to why Intel 18A is garnering [[link]] such attention, there are two important reasons. Firstly, the process node is claimed to be roughly on par with TSMC's N2—a little worse in terms of transistor density but better in performance, but close enough to potentially be viable for key products, especially monster GPUs for AI servers.
Naturally, the report doesn't mention any specific details or sources for the information, but it should come as no surprise that AMD, Broadcom, and Nvidia are evaluating a chip manufacturer, as they have no such facilities to do so themselves. With the current options being TSMC or Samsung, the processor market is heavily reliant on a small supply chain—having one more viable option will not only keep investors happy, but could help to ensure processor stocks can meet demand.
