Ever seen a person eat his personal head? In all probability not, however you are very doubtless going to see one thing nearly as unusual, an Nvidia GPU made by none apart from arch rival Intel.
Nvidia’s CEO and leather-based jacket aficionado Jensen Huang has confirmed that the corporate plans to make use of Intel’s new cutting-edge chip manufacturing facility in Arizona (opens in new tab).
At this level Nvidia is not detailing what chips precisely. So, it might be one thing apart from GPUs. However Nvidia is primarily a GPU firm, so Nvidia GPUs made by Intel simply grew to become a really reasonable chance.
After all, we have seen oddities like this earlier than, as arch rivals come collectively in some pragmatic shared trigger. It wasn’t that way back that Intel teamed up with AMD to insert Radeon Vega M graphics into an Intel cell CPU package deal (opens in new tab).
That by no means actually appeared to catch on and Intel has since embarked by itself graphics chips with the Arc collection of GPUs (opens in new tab). Nevertheless it’s an instance of how a aggressive relationship will also be symbiotic.
Making chips for patrons is central to Intel’s plan to reinvent itself as not solely a full built-in chip design and manufacturing powerhouse, but in addition as a foundry to compete with the mighty TSMC.
As we not too long ago reported, Intel is sticking publicly to its aggressive plans to roll out no fewer than three new chip manufacturing nodes within the area of a yr (opens in new tab), beginning with its new Intel 4 node, previously often known as 7nm.
Intel 20A supposedly follows in early 2024 with Intel 18A in late 2024, at which level Intel claims it can have probably the most superior chip manufacturing expertise round.
For Nvidia, a leading edge various to TSMC could be very helpful certainly. Partly, that may assist to maintain TSMC trustworthy on the subject of pricing. However in strategic phrases, being a fabless firm counting on only one possibility for manufacturing should be a bit terrifying, particularly when that possibility is positioned on an island within the South China Sea that’s at present the topic of utmost tense geopolitics.
Admittedly, Nvidia did go to Samsung for its Ampere technology of GPUs together with the RTX 3090 (opens in new tab). So TSMC is not actually the one possibility. However Ampere, although pretty efficient in efficiency phrases, was a reasonably inefficient structure, in no small half because of the traits of Samsung’s 8nm node, which most trade observers take into account far inferior to TSCM’s 7nm expertise of the time.
Anyway, it isn’t a nailed on certainty that Nvidia GPUs shall be made by Intel. Heck, it is from assured that Intel will get its manufacturing expertise sufficiently up to the mark that Nvidia will even wish to use it. However Nvidia graphics made by Intel very a lot appears to be the plan. And that may take some getting used to.
