Intel (INTC.US) shares fell sharply in pre-market trading on Wednesday, dropping over 5%, following reports that NVIDIA (NVDA.US) has suspended testing of Intel’s 18A advanced process technology after an initial evaluation. According to two sources familiar with the matter, NVIDIA, the dominant player in AI chips, decided not to proceed further with testing Intel’s 18A process for manufacturing one of its AI chips.
Earlier in March, media reports indicated that NVIDIA and Broadcom (AVGO.US)—two leading fabless AI chip designers—had engaged in high-end chip manufacturing tests with Intel. However, Intel’s 18A technology has since faced challenges, with initial test results disappointing clients and potentially pushing them to rely on TSMC (TSM.US), Intel’s primary competitor in semiconductor manufacturing.
For Intel, securing major clients is critical to its foundry ambitions. The company is aggressively investing in cutting-edge process technology, including the acquisition of High-NA EUV lithography machines—the first of their kind from ASML—to develop 2nm and more advanced chipmaking capabilities, aiming to surpass TSMC and Samsung.
While Intel has rebounded from its downturn thanks to strong server CPU demand, shareholders expect sustained growth, positioning the company as a foundry giant rivaling TSMC. The success of its 18A and upcoming 16A processes is pivotal to this strategy. The 18A node (equivalent to 1.8nm-class technology) incorporates RibbonFET transistors and PowerVia backside power delivery to enhance performance and efficiency, making it a key offering for both Intel’s own products and external foundry clients.
NVIDIA’s decision to pause testing reflects a cautious supply chain strategy rather than a complete rejection of Intel’s technology. In September, NVIDIA CEO Jensen Huang announced a $5 billion investment in Intel, signaling a rare collaboration to develop high-performance chips for PCs and AI data centers. The news fueled a 50% surge in Intel’s stock since then, raising hopes for future AI chip orders from NVIDIA.
As the world’s largest AI GPU designer, NVIDIA demands stringent performance, power efficiency, and yield from its manufacturing partners. The suspension suggests NVIDIA is prioritizing stability and performance by sticking with established partners like TSMC. However, this does not rule out future collaboration—NVIDIA had previously participated in early 18A testing and may revisit the process once yields improve.
Comments