Friday, August 21st 2026
Intel "Nova Lake-S" 28-Core SKU With bLLC to Draw Nearly 300 W at PL2
Intel's upcoming "Nova Lake-S" desktop processors are nearing launch, and the rumor mill is intensifying as we receive more information about clock speeds, core configurations, and power consumption. According to a post on the Weibo forums, Intel's 28-core SKU with a big last-level cache (bLLC) integrated into the package will reach up to 296 W of power consumption at the Power Level 2 (PL2) setting. This is the maximum power draw the CPU can achieve during short bursts of turbo clocking, reaching its peak. Compared to the current top-end "Arrow Lake-S" 24-core Core Ultra 9 285K, which has a PL2 setting of 250 W, the additional four LPE-Cores on the latest "Nova Lake-S" will increase that number by about 18.4%.
Previously, we learned that "Nova Lake-S" will be very power-hungry, especially in the top-end SKUs. For example, Intel's top-end, overclocking-focused, dual compute die "Nova Lake-S" CPU with 52 cores will feature a PL2 mode with a power level of 474 W. This is an unprecedented number for an Intel consumer desktop platform, typically expected from a high-end desktop (HEDT) design. However, since this is a 52-core design that can be overclocked, it represents the highest CPU PL2 setting that Intel will be shipping with "Nova Lake-S." For single compute die SKUs, the PL2 limit appears to be 296 W.Finally, this could be the previously leaked Intel Core Ultra 9 4950K, identified as a 28-core "Nova Lake-S" desktop chip. It features 8 "Coyote Cove" P-Cores, 16 "Arctic Wolf" E-Cores, and 4 LPE-Cores. Intel is testing many SKUs for "Nova Lake-S," and a four-digit naming scheme could be necessary due to the numerous SKUs and configurations, so the new naming scheme might be on the way.
Source:
HXL on X
Previously, we learned that "Nova Lake-S" will be very power-hungry, especially in the top-end SKUs. For example, Intel's top-end, overclocking-focused, dual compute die "Nova Lake-S" CPU with 52 cores will feature a PL2 mode with a power level of 474 W. This is an unprecedented number for an Intel consumer desktop platform, typically expected from a high-end desktop (HEDT) design. However, since this is a 52-core design that can be overclocked, it represents the highest CPU PL2 setting that Intel will be shipping with "Nova Lake-S." For single compute die SKUs, the PL2 limit appears to be 296 W.Finally, this could be the previously leaked Intel Core Ultra 9 4950K, identified as a 28-core "Nova Lake-S" desktop chip. It features 8 "Coyote Cove" P-Cores, 16 "Arctic Wolf" E-Cores, and 4 LPE-Cores. Intel is testing many SKUs for "Nova Lake-S," and a four-digit naming scheme could be necessary due to the numerous SKUs and configurations, so the new naming scheme might be on the way.

125 Comments on Intel "Nova Lake-S" 28-Core SKU With bLLC to Draw Nearly 300 W at PL2
tpucdn.com/review/intel-core-ultra-7-270k-plus/images/power-applications-compare-core-i9-14900k.png
Well then good luck Intel, you'll need it.So the LPE cores are active during load? Why?
I thought they are only active in idle and very low load while the compute tile is disengaged, after the compute tile is engaged why do those little turds keep being active? For 1 extra fps and 1 extra CB point?
I hope AMD steamrolls these amateurs so I can buy this crap at a big discount and then put a nice 200 W power limit on it. :nutkick:
Also, cars analogies are hard to make in a broad sense here. No, the guy buying a supercar isn't worried about fuel economy, but the guy buying a commuter vehicle certainly would. Even then, automobile purchases are largely an emotional decision.Yeah. I mean, we forum folk are the ones who often see or experience the product failures the most, either personally or through support. Besides, it's not like we haven't seen motherboard manufacturers abuse power curves to the detriment of specific SKUs (X3D) or entire product stacks (RPL).
Arrow lake draws 250W and in multithreading performance it is on par with the most powerful CPUs from AMD. which draw similar power. So your point was ... ???
Note now far AMD pushes the poor 8 core CPU:
www.techpowerup.com/review/amd-ryzen-7-9850x3d/23.html
This is truly a torture of the chip. :(
I'm so glad I bought an Arrow Lake 270K on sale a couple of months ago. I knew it wouldn't be worth waiting for Nova Lake.
Single tile/CCD configurations should max out at ~150W and dual tile/CCD configurations should max out at ~300W.
If consumers want to increase power limits over that, don't stop them. But in a nicer world they wouldn't ship with such absurd power limits.
All this to say, this is why such an analogy is hard to make here. Though I guess my wife requires a keypad on her laptops, which eliminates many quality options. However, it puts a seemingly superfluous spec ahead of things like CPU or GPU performance, which only need to be adequate to her.
Back in Alder Lake days, PL2 value itself was not the issue. The issue was that Intel's guidance allowed motherboard manufacturers to basically run their mobos in PL2 mode all the time. Has this been corrected in Arrow Lake? Will it be corrected in Nova Lake?
I mean, I did the BIOS changes manually, it's not a big deal, and my 12600k has always run cool and quiet (yes, I went there). But it would be nicer to get a CPU that behaves out of the box. Not for me, but for many others.
Over 400w 285K
Almost 500w 14900K
Arrow lake has no such issues, and for the most part intel chips handle high wattages just fine, and you can easily undervolt them - same as AMD these days. Undervoling and OCing to get some extra performance out of them is the way to go in both camps.
The point is - there is no sweet spot. There is no optimal power draw. The more you limit it, the more efficient it gets. You can see that the energy consumed to perform the defined task drops depending on how much limited the CPU is.
So if they sell a CPU with a high power limit, that is set in such a way that the CPU survives running some typical load for the time somebody defined to be the life of the CPU, it is up to the user to limit the CPU how they want, if they want better efficiency, if they want to run extremely demanding tasks 24/7, if they are limited by the cooler performance, etc.
You can see you save almost half of the energy when power limiting the 270K to 100W and lose only 20% of the performance.
So if 270K is an extreme inefficent version of the CPU, with what power limit should 270 non K sell, if it existed, to that it is a nice cool running enviroment friendly efficient product? Is it 100W? 125? 150? I am not sure, but it certainly should be somehow limited compared to the extreme K model.
The LPE cores being active during load creates a countermeasure to the compute tile hotspot.
This way the heat is more spread out, there are more watts than without LPE cores active but the thermal density (W/active area) is less thus it will be easier to cool.
One other aspect could be unaffected 100% performance of the compute tile in heavy MT tasks because the LPE cores are busy keeping the OS bloat and background BS in check, otherwise those would take away performance from the compute tile in heavy MT tasks.
Sort of like "yo dawg we put a CPU in your CPU so you (actually the OS in the background) can compute while you compute".100% because they don't have any Zen6 to compare so they have to beat current Zen5 X3D with enough margin so that they're confident it can hang with Zen6 X3D.The 9950X3D almost gets to 215 W in Keyshot. Clearly less than the 260 W for 270K/285K.
tpucdn.com/review/amd-ryzen-9-9950x3d2-dual-edition/images/power-applications-compare-ryzen-9-9950x3d.png
tpucdn.com/review/intel-core-ultra-7-270k-plus/images/power-applications-compare-core-ultra-9-285k.png
www.techpowerup.com/327004/intel-isolates-raptor-lake-vmin-shift-instability-root-cause-new-microcode-update-coming