Higher volume manufacturing with the early yield rates of Intel 18A has been reported as not economically viable, pushing the possible entry into 2026. Intel has dismissed its HVM plans for 18A and aims to work on Intel 18A yields before scaling as different industry reports indicate Intel 18A yield rates are currently far below mass production needs.
Intel 18A–which Panther Lake laptop CPUs were expected to run on, and a flagship for Intel’s foundry ambitions. Panther Lake timelines may be in for a delay due to Intel 18A HVM being hedged, as will the external foundry push of Intel 18A, which is also being revisited. Intel 18A yields have been measured in the single digits to low double digits percentages in some tests, making ramping Intel 18A potentially dangerous for high volume manufacturing, according to reports.
Early reports differ on the figures, but the main risk is fairly clear: Intel 18A requires better repeatable yields ahead of HVM. Intel 18A yield uncertainty has already had corporate leaders evaluating the trade-offs between getting out on-time for a 2025 product and preserving margins and product quality.
Sources say Intel 18A yields are not acceptable for the company to be successful and its leadership is apparently not prepared to accept inferior 18A yields. According to IDM 2.0 CEO Lip-Bu Tan and Intel leadership is reportedly considering whether to aggressively market Intel 18A to external customers while yields are still unstable. INTC may now choose to delay Intel 18A in favor of better Intel 18A quality while protecting margins.
Delay of Intel 18A is strategic upside
For some, it might even be a tactical reason to place Intel 18A on the back burner. This gives intel an extra amount of time to improve Intel 18A yields as early HVM of a complex node ususally leads to losses which Intel can avoid. If Intel’s 18A process at scale is indeed stronger, that would bring Intel competitive with TSMC, and would also void the prospects for Intel’s foundry business. A delayed but superior Intel 18A may ensure more stable production and generate fewer recalls or operating losses, which bodes well for investors and partners.
In short, it deprives Panther Lake of any chance of making the sizable performance gains that Intel is planning for the disaggregated SOC that Panther Lake is based on and would also appear to be the last nail in the coffin for any truly competitive, standalone foundry plans from Intel at least through 2021.
Intel is still fine-tuning Intel 18A production, so Panther Lake’s trip to market may be further modified. If Intel’s 18A schedule slips, it could push internal product schedules, but just as likely it ultimately delivers a stronger Panther Lake on a stronger, mature Intel 18A node. And Intel’s wider foundry efforts rest on yielding competitive Intel 18A that lure external customers.
Conclusion
On the surface, Intel 18A being pushed back to 2026 sounds bad, but it could indicate a no-excuses, quality-first mentality. Better Intel 18A yields prior to HVM can protect shareholder value and re-establish Intel’s foundry for a stronger recovery. With timings, timeline, yield improvements, and other strategic choices to balance, Intel 18A developments are one to not take your eyes off of just yet.
FAQ
Sources say Intel 18A yield rates are too low for high-volume manufacturing to be profitably high, so the helter-skelter HVM postponement gives Intel the time to perfect Intel 18A yields and protect margins.
Not necessarily. Or, with the caveat that Intel targets Panther Lake to ship on a more mature Intel 18A node, which would probably necessitate timeline adjustments for Panther lake while it constructs a more 18A production.
Leadership at Intel is said to be evaluating the way and time of the external promotion of Intel 18A; the final approach may rely on process yield capabilities and business case.
Different reports give different public figures; many cite low yields (i.e. single / low double digit or ~10% in some accounts). Intel did not provide precise public yield figures.
