Wednesday, June 18th 2025
Intel Showcases 18A Node Performance: 25% Faster and 40% Lower Power Draw
Intel's presentation at the VLSI Symposium in Japan offered a detailed look at the upcoming Intel 18A process, which is set to enter mass production in the second half of 2025. This node combines Gate-All-Around transistors with the PowerVia backside power delivery network, resulting in a completely new metal stack architecture. By routing power through the rear of the die, Intel has been able to tighten interconnect pitches on critical layers while relaxing spacing on the top layer, improving yield and simplifying fabrication. In standardized power, performance, and area tests on an Arm core sub-block, Intel 18A demonstrated roughly 15% higher performance at the same power draw compared to Intel 3. When operating at 1.1 volts, clock speeds increase by up to 25% without incurring additional energy costs, and at around 0.75 volts, performance can rise by 18%, or power consumption can drop by nearly 40%.
Under the hood, the process features significant cell height reductions: performance-tuned cells measure 180 nanometers tall, while high-density designs sit at 160 nanometers, both smaller than their predecessors. The front-side metal layers have been reduced from between 12 and 19 on Intel 3 to between 11 and 16 on Intel 18A, with three additional rear metal layers added for PowerVia support. Pitches on layers M1 through M10 have been tightened from as much as 60 nanometers down to 32 nanometers before easing again in the upper layers. Low-NA EUV exposure is used on layers M0 through M4, cutting the number of masks required by 44% and simplifying the manufacturing flow. Intel plans to debut 18A in its low-power "Panther Lake" compute chiplet and the efficiency-core-only Clearwater Forest Xeon 7 family. A cost-optimized 17-layer variant, a balanced 21-layer option, and a performance-focused 22-layer configuration will address different market segments.
Source:
via HardwareLuxx
Under the hood, the process features significant cell height reductions: performance-tuned cells measure 180 nanometers tall, while high-density designs sit at 160 nanometers, both smaller than their predecessors. The front-side metal layers have been reduced from between 12 and 19 on Intel 3 to between 11 and 16 on Intel 18A, with three additional rear metal layers added for PowerVia support. Pitches on layers M1 through M10 have been tightened from as much as 60 nanometers down to 32 nanometers before easing again in the upper layers. Low-NA EUV exposure is used on layers M0 through M4, cutting the number of masks required by 44% and simplifying the manufacturing flow. Intel plans to debut 18A in its low-power "Panther Lake" compute chiplet and the efficiency-core-only Clearwater Forest Xeon 7 family. A cost-optimized 17-layer variant, a balanced 21-layer option, and a performance-focused 22-layer configuration will address different market segments.




25 Comments on Intel Showcases 18A Node Performance: 25% Faster and 40% Lower Power Draw
I wonder if they have something that can combine sram and foveros so they can do their own version of x3d
ryzen first models were ok at best.. but it got much better...
we just need intel next product to be faster at productivity/gaming. runs fairly cool and not degrade out of the box that requires microcode to fix...
TSMC's 2nm process is far superior in regard to yields. Even Intel has ordered some volumes in TSMC's 2nm production queue.
I agree that Intel needs another moment. More competition, the better.
However, back on topic, we do need competition to improve in the semiconductor fab space as TSMC is the only real player. This is why I keep hoping that Intel will pivot away from chip design and concentrate only on a fab for hire model. Intel can't do both and will just go out of business if they try.
I keep wondering why memory latency is that high, after all. People (seemingly dumb people) tend to bemoan that Intel has moved the memory controller away from the cores, and there was much appreciation when they announced they had moved them back, but … Look how AMD’s chiplet chips need to traverse the package (!) to reach the memory controller from the cores, so that really can’t be where the excess latency comes from.
I have now re-read more on that than I had wished for, and … is it even sure, now, what’s the problem? All I could see is that L3 is also slow, and that’s already part of why it takes so long to get to memory—even though L3 is where it’s always been.What kind of stuff? Positive news about their process? How dare they! :rolleyes: FWIW, the VLSI Symposium is a long-standing industry event, where it’s tradition to share that exact kind of info, so nohting out of the normal there.You say ‘even’, yet I don’t see any connection between these statements.
End users however need competition and Intel needs to get their s*** together with this 18A node or the only Intel there will be, is the one in history books.
AMD slashed their prices in half overnight at its launch, and never recovered until Ryzen.Yes it was.
In short: Show, don't tell, Intel. Let's see actual product and then we'll judge...
Imagine if we had 64TB SATA drives which is possible, for the price of say a 12TB HDD. Ironically as you said they would make as much money from the massive higher amount of sales.
Besides, Intel has only ever once outsourced the production of their CPU throughout the half century of selling CPUs, and that was with Lunar Lake using TSMC N3. I don't think that a single occurrence constitutes a pattern that justifies reacting like that.