Intel used the 2026 VLSI Symposium in Honolulu this week to confirm that its 18A-P process node has officially entered risk production, hitting the timeline it promised customers and partners last year. That is a meaningful checkpoint. Risk production means the node is transitioning from R&D to early-stage manufacturing, with data indicating it will meet customer requirements before full qualification.

The numbers Intel put on the board are solid for what is a node refresh rather than a full generational leap. It looks like 18A-P delivers 9% higher performance at the same power draw compared to standard 18A, or 18% lower power consumption at equivalent performance. Thermal resistance improves by 20-40%, and via resistance drops by 10-30%.

In terms of features, Intel also unveiled Power Boost, a new dual-contact, low-resistance transistor option that increases drive current and gives designers more frequency headroom. As we previously covered, these figures were already circulating before VLSI, and Intel has now officially confirmed them all.
- Read more: Intel's new 18A-P node will reportedly provide 9% higher performance and 18% better efficiency than 18A
- Read more: Intel's 14A node will enter risk production in 2028, while 10A and 7A nodes are on the roadmap
- Read more: Intel introduces Xeon 7 'Diamond Rapids' CPU lineup, built on the 18A-P process node

The first major product landing on 18A-P will be Xeon 7 "Diamond Rapids", confirmed for a 2027 launch. We previously covered reports that those chips could pack 192 Panther Cove-X P-cores across four compute tiles, 16-channel DDR5, and PCIe Gen 6 support. AMD's EPYC Venice is still on track to ship this year with up to 256 Zen 6 cores, so Intel will be playing catch-up in raw core count, but a credible process node gives it something real to lean on.

Intel Foundry has been bleeding money for years, and 18A-P entering risk production on schedule is exactly the kind of proof of execution that external customers need to see. Apple is reportedly in talks to use 18A-P for its M7 SoC, targeting MacBook Air and iPad Pro volumes in the 15-20 million unit range annually.
Microsoft has also been linked to the node for its next-gen Maia AI accelerator, and Google is reportedly exploring Intel's advanced packaging options. The US government is already a paying customer, and Tesla is signed on for the 14A node that follows. Intel CEO Lip-Bu Tan has publicly flagged that he expects multiple foundry commitments to close in the second half of 2026. Hitting production milestones on schedule is how he gets there.
The foundry is projected to break even in 2027 if those deals land, with the 14A node also expected to enter risk production in 2028. This VLSI announcement is Intel showing it can execute on its plans. The question now is whether the yields hold up at volume.




