Why NVIDIA is skipping Intel 18A for its next generation AI chips
NVIDIA tested Intel Foundry’s 18A node but has not committed to mass production, Reuters reported. Two individuals familiar with the matter stated NVIDIA recently evaluated manufacturing chips using Intel’s 18A production process but did not proceed further. NVIDIA did not respond to a request for comment, according to Reuters.
Numerous customers have approached Intel for 18A node evaluations, a common practice for chip designers assessing competitors to Taiwan Semiconductor Manufacturing Company (TSMC) offerings before committing to additional foundry capacity. The 18A node primarily serves Intel’s internal products; however, future versions, including 18A-P and 18A-PT, aim to attract external customers and are slated as long-term Intel Foundry products.
Intel Foundry’s 14A node is undergoing active development toward a 2027 release and is becoming a critical product. Intel is collaborating with customers on the node design, with potential foundry partners assessing its suitability. Early feedback indicates customer satisfaction with development progress, describing the node as competitive.
An Intel-NVIDIA partnership formed in mid-September is anticipated to manifest starting in 2026, integrating NVIDIA RTX GPUs within Intel x86 SoCs. The agreement currently focuses on product delivery, and it remains unclear if it will involve foundry services for NVIDIA. Further feedback from potential foundry customers is expected upon the release of the 14A 0.5 PDK. The 14A node is set to debut in 2027 alongside the 18A-P and 18A-PT for specialized applications.