NVIDIA is reportedly exploring options to diversify its chip supply chain, and Intel Foundry is an optimistic choice for next-gen AI lineups at the moment.
NVIDIA Is Looking at a "Low Risk" Partnership With Intel Foundry, Involving the Outsourcing of Just the Feynman I/O Die
TSMC is emerging as a major bottleneck in the AI supply chain, and there's no doubt that fabless manufacturers are exploring options to expand their sourcing strategies. With the recent breakthroughs Intel has seen with its 18A process, the foundry division has managed to bring the momentum that could potentially lead to external customers flowing in, and according to a report by DigiTimes, NVIDIA is now exploring Intel's 14A/18A process for Feynman's I/O die, along with EMIB advanced packaging.
It is claimed that NVIDIA won't go "all in" on Intel Foundry for Feynman and will instead use chips from TSMC as well. The report says that Team Green will outsource only a Feynman die to Intel and will likely adopt either the 18A or the 14A node, though the latter seems more likely. Moreover, Intel will also provide its EMIB advanced packaging for Feynman, and in terms of the manufacturing split, Intel will get 25% of the entire Feynman production, while the rest will be allocated to TSMC.
Several US fabless manufacturers are aggressively pursuing a dual-foundry strategy, with AMD and Qualcomm also looking to pair Samsung and TSMC. This pursuit is driven by two key bottlenecks: the first is that the entire AI supply chain currently depends on the Taiwan giant for front-end and back-end needs, which creates a 'single point of failure' in case geopolitical rivalry between China and Taiwan heats up. Secondly, because the AI buildout is so massive, manufacturers and hyperscalers are scrambling to secure capacity at TSMC, yet many are unable to do so.
NVIDIA's approach with Feynman is a "low risk" one, as assigning the I/O die to Intel Foundry ensures volume ramp-up isn't compromised if Intel fails to deliver on yield or capacity. And more importantly, NVIDIA is also looking to outsource 'non-core' products to Intel Foundry, which means next-gen gaming GPUs could be part of a foundry deal with Team Blue.
It would be interesting to see how the situation at Intel Foundry evolves going forward, since, by the looks of it, TSMC's customers are now forced to diversify their supply networks, leaving only Samsung and Intel as viable options.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
