Recently, rumors surfaced that NVIDIA has ceased testing Intel's 18A process, a development that sent Intel's stock price tumbling. The market perceived this as a setback for Intel's ambitions to bolster its foundry business. Yet, those well-versed in Intel's foundry endeavors should recognize that, irrespective of the news's veracity, it's improbable that NVIDIA would currently opt for Intel's process to fabricate its GPUs and other chips.

NVIDIA had, in fact, previously trialed Intel's 18A process but chose not to advance the partnership further. This decision underscores that the process still harbored flaws and failed to align with NVIDIA's stringent standards for AI chip production. Although Intel's 18A process incorporates a slew of innovative technologies, including RibbonFET transistors and PowerVia backside power delivery technology, yield challenges persist as a pivotal hurdle impeding its widespread adoption.

Presently, Intel is in the process of developing a low-power iteration of the 18AP process and has outlined plans to unveil the 14A process, which boasts greater market potential externally, to entice external clientele. For Intel, the road ahead in expanding its foundry business is fraught with challenges. It must enhance process yield and performance, while also navigating the competitive dynamics with potential customers in the chip industry.