Intel says its next chip-making process will land "within 5%" of TSMC's competing node - which reads like either a promise or a warning, depending on which way that 5% goes.
Naga Chandrasekaran, chief technology and operations officer of Intel Foundry, told analysts at KeyBanc that the company's upcoming 14A process is expected to perform "within 5%" of TSMC's A14, according to a report citing the note. Both processes target the 1.4-nanometer class and are still years from shipping. Here's the catch: Intel's own public roadmap claims 14A will deliver 15-20% higher performance than its current 18A node at the same power, while TSMC pegs A14's gain over its N2 node at 10-15%. Run Intel's own numbers and 14A should be comfortably ahead of A14, not neck-and-neck.
That gap between what Intel's roadmap implies and what its foundry chief is now telling investors is the real story. Foundries sell confidence to customers picking a manufacturing partner for the next decade of chips, so a leader hedging its own numbers is unusual - worth reading as either candor or cover for a process that's tracking behind its targets.
Intel has missed process promises before, and CEO Lip-Bu Tan has made a habit of setting expectations low. "Within 5%" might just be Intel doing that again, one node early.