WTF?! Could your future Nvidia GPU be produced by an Intel foundry? As unlikely as that sounds, staff inexperienced boss Jensen Huang says it is one thing his firm is contemplating. Intel boss Pat Gelsinger responded to the information by saying he was “thrilled” for Nvidia’s curiosity, confirming that the 2 firms have been engaged in ongoing discussions.
Huang answered questions at a press Q&A session yesterday (through Tom’s Hardware) and was requested about Intel’s Foundry Services (IFS) initiative. “We are very open-minded to contemplating Intel, and I’m delighted by the efforts that they are making,” he mentioned.
Intel and Nvidia compete within the CPU And GPU markets so the revelation could appear somewhat unusual, however Haung defined that “Our technique is to increase our provide base with range and redundancy at each single layer. At the chip layer, on the substrate layer, the system layer, at each single layer. We’ve diversified the variety of nodes, we have diversified the variety of foundries, and Intel is a superb accomplice of ours […] They’re inquisitive about us utilizing their foundries, and we’re very inquisitive about exploring it.”
Intel CEO Pat Gelsinger instructed Reuters that the corporate is “thrilled for his or her [Nvidia] curiosity in utilizing our foundry capabilities.” He confirmed that discussions have been ongoing, however there have been no stable timelines—Huang added that foundry talks can take a very long time.
“it is not nearly want,” Huang mentioned. “We should align know-how, the enterprise fashions should be aligned, the capability must be aligned, the operations course of and the character of the 2 firms should be aligned. It takes a good period of time and loads of deep, deep dialogue – we’re not shopping for milk right here. This is de facto in regards to the integration of the availability chains. Our partnerships with TSMC and Samsung within the final a number of years are one thing that took years to domesticate. So we’re very open-minded to contemplating Intel, and I’m delighted by the efforts that they are making.”
Most firms can be hesitant to work so carefully with rivals in case it exposes commerce secrets and techniques, however Huang is not frightened: he says each Intel and AMD have recognized Nvidia’s secrets and techniques for years, and there are many authorized methods to maintain habits on the extent, resembling NDAs. “We share roadmaps, after all, below confidentiality and a really selective channel of communications. The trade has simply discovered work in that approach.”
Speaking about tech rivals collaborating in mutually helpful methods, Huang cited some examples of how his firm has labored with AMD and Intel, together with the previous’s CPUs showing in DGX methods. “If not for Intel CPUs and the entire hyperscalers linked to our HGX, we would not be capable to ship HGX,” Huang mentioned, including that Intel processors are utilized in Nvidia’s Omniverse platform.
Nvidia is presently centered on producing this yr’s RTX 4000 collection, aka Lovelace, for which it has reportedly paid TSMC $10 billion to safe a big chunk of the Taiwanese agency’s 5nm manufacturing capability. The firm presently makes use of Samsung’s 8nm course of node to make the buyer Ampere GPU line.