Intel's 14A (1.4-nanometer-class) process node is seeing defect density reduction that is outpacing internal expectations, marking the company's best performance since its 22nm node. CFO David Zinsner revealed at Deutsche Bank's 2026 Technology Conference that "we haven't seen performance like this since 22nm."
The 14A node is scheduled for risk production in the second half of 2027, with high-volume manufacturing targeted for 2028. With roughly two years until production begins, tangible progress is already emerging on the customer front: Intel's internal design teams have begun developing products on 14A, while external foundry customers have shifted from evaluating technical data to inquiring about capacity allocation.
Drawing Parallels to 22nm: Promising Signs with Three Important Caveats
"14A's defect density is better than our target curve, and the rate of defect reduction is faster than any previous node. We haven't seen this kind of performance since 22nm—and 22nm was arguably one of the best process nodes Intel has ever delivered," Zinsner stated during the conference.
The 22nm node was Intel's first FinFET process and debuted well ahead of the industry, with competitors only following suit at their 14/16nm nodes between 2014 and 2015. After the Ivy Bridge processor launched successfully in 2012, 22nm served Intel's CPU product lines through 2016 and continued powering other product families afterward.
However, Zinsner's comparison comes with strict boundaries. He is benchmarking 14A's defect reduction trajectory at roughly two years before production against where 22nm stood at a similar stage around 2010, rather than making a direct comparison of absolute defect levels between the two generations. Additionally, advances in wafer inspection equipment over the past 16 years have changed how "defects" are measured, and defect density itself does not directly translate to yield.
The technical complexity of 14A makes these results even more notable: second-generation gate-all-around (GAA) RibbonFET transistors, second-generation backside power delivery (PowerDirect), and High-NA EUV lithography present far greater challenges than 22nm ever posed. Achieving rapid defect reduction under such a demanding process combination stands out strikingly in Intel's recent process node history—where 14nm faced a year-long delay due to yield issues, the original 10nm node was deemed a failure, 20A was cancelled, and 18A still carried elevated defect levels at production, while Intel 4 and Intel 3 saw almost no public progress updates.
Customer Sentiment Shifts from Data Reviews to Capacity Demands
Zinsner revealed that Intel's internal design teams are already developing products on 14A. He described these internal clients as "probably the most demanding of all," adding that "their decision to design products on 14A is a significant vote of confidence for us."
External foundry customers are also showing a qualitative change in attitude. According to Zinsner, CEO Lip-Bu Tan and his team currently meet with customers on a weekly basis, noting that "customers have moved from reviewing data to asking 'How much capacity can I get? What's the supply timeline?' We now have conviction about external customers for 14A as well."
With roughly two years still remaining before high-volume manufacturing, Intel expressed confidence that the 14A node has the potential to become a long-lifecycle process in the same vein as 22nm.
Comments