Wednesday, February 11th 2026
Intel "Nova Lake" Compute Tile Die-sizes Surface
Intel is preparing to launch its next generation Core Ultra Series 4 "Nova Lake" processors in the second half of 2026. We have some hint on what the sizes of the Compute tiles of these chips could measure. "Nova Lake" remains a tile-based disaggregated chip, just like the current "Arrow Lake," with the most advanced foundry node being allocated only to the components that can benefit the most from it, the CPU cores. The CPU complex of the processor is located in the Compute tile, while low-power island E-cores are located on the SoC tile that's built on a slightly older foundry node.
Intel is building the Compute tile of "Nova Lake" on the TSMC N2 (2 nm nanosheet) foundry node. There will be different types of Compute tiles based on the CPU core count and cache sizes. The two most popular ones for the desktop platform will be the mainstream tile with 8P+16E core configuration and a standard L3 cache size shared among the eight P-cores and four E-core clusters. The second, more premium Compute tile type will be 8P+16E with bLLC, or big last-level cache. bLLC is an enlarged L3 cache that's estimated to be 3-4 times the size of the regular L3 cache. This is Intel's answer to AMD's 3D V-Cache. Estimates put the regular Compute tile to measure around 110 mm², while the premium variant with bLLC is estimated to measure over 150 mm². This 36% increase in die-area is entirely from the enlarged last-level cache.This isn't the entirety of the "Nova Lake" chip, there will be other components, such as the SoC tile that contains the low-power island E-cores, the DDR5 memory controllers, the 74 TOPS NPU, and the PCI-Express Gen 5 root complex, besides the System Agent, Management Engine, and security processor. Besides the SoC tile, are the Graphics tile, which is likely to be built on a more advanced node than the SoC, but less advanced than the Compute tile. Intel has plenty of options for the Graphics tile, including its own Intel 4 foundry node.
Source:
HXL (Twitter)
Intel is building the Compute tile of "Nova Lake" on the TSMC N2 (2 nm nanosheet) foundry node. There will be different types of Compute tiles based on the CPU core count and cache sizes. The two most popular ones for the desktop platform will be the mainstream tile with 8P+16E core configuration and a standard L3 cache size shared among the eight P-cores and four E-core clusters. The second, more premium Compute tile type will be 8P+16E with bLLC, or big last-level cache. bLLC is an enlarged L3 cache that's estimated to be 3-4 times the size of the regular L3 cache. This is Intel's answer to AMD's 3D V-Cache. Estimates put the regular Compute tile to measure around 110 mm², while the premium variant with bLLC is estimated to measure over 150 mm². This 36% increase in die-area is entirely from the enlarged last-level cache.This isn't the entirety of the "Nova Lake" chip, there will be other components, such as the SoC tile that contains the low-power island E-cores, the DDR5 memory controllers, the 74 TOPS NPU, and the PCI-Express Gen 5 root complex, besides the System Agent, Management Engine, and security processor. Besides the SoC tile, are the Graphics tile, which is likely to be built on a more advanced node than the SoC, but less advanced than the Compute tile. Intel has plenty of options for the Graphics tile, including its own Intel 4 foundry node.

55 Comments on Intel "Nova Lake" Compute Tile Die-sizes Surface
Nova Lake won't have a price advantage because of a new socket, and I expect Intel will price a 52 core CPU towards the very high end market.
oh my bad edited 3 MB instead of 1.5.
www.techpowerup.com/336412/inside-arrow-lake-intels-die-exposed-and-annotated
The L3 cache has doubled that's true but given the smaller node I would've expected a larger difference.I doubt Zen6 vanilla matches let alone beats Zen5 X3D in gaming.
Right Arrow Lake is close to Zen5 vanilla in gaming, that means Nova Lake vanilla can be pretty close to Zen6 vanilla, it's not unrealistic.
If the bLLC Nova Lake can be just as close to Zen6 X3D remains to be seen but at worst I'd say it will match Zen5 X3D.Yes but that is with pretty much equal frequency. If Zen6 has significantly higher frequency that can offset quite a bit of the thread advantage.Yes, for the Zen5 crowd.
What about the people on LGA1700, LGA1851, AM4 or older platforms?
If they want to upgrade whichever side they choose they have to buy the full platform. True that some of those on LGA1700 and those on LGA1851 can reuse their DDR5 RAM.
It will be a battle to win these people over so both CPU makers have to work hard for it.
LGA1700, or AM4 I can see, but anyone with LGA1851 I wouldn't recommend unless Nova Lake will be a significant upgrade.
Building a PC isn't very enjoyable unless an upgrade is needed because of skyrocketed RAM & SSD cost, both AMD and Intel competing for price will be interesting to see.
You’re welcome.Thank you for that incredible insight. How many hours of graduate level research did you have to put in to come to the determination that 52 CPU cores are expensive to produce?
Zen 5 failed to impress against Zen 4 X3D, but that was true of vanilla Zen 5 universally. At least in consumer workloads.
The problem for Intel and gaming is price.
Even if we assume their top end 485K or whatever it ends up being called beats Zen 5 X3D with a decent margin and gives Zen 6 a run for it's money - it's still presumably a $700 chip. AMD will release Zen 6 gaming parts (single-CCD) below 500 and Zen 6 X3D presumably at 500.
Even if Intel somehow claims the throne it's a pyrrhic victory with much more expensive and power hungry part.
That means Intel's midrange 465K and 445K with reduced amount of bLLC wont be as competitive in gaming performance, even if they are competitive in price.
5800X3D: $449
7800X3D: $449
9800X3D: $479
9850X3D: $499
Over the course of 3+ years and two new generations the top single-chiplet X3D price has risen by $50. And all of these were popular.
I fail to see how in two years with one new generation this price now suddenly jumps to $579 when the competition is closer than it has been in years?
I predict flat prices. Tho TSMC N2 is more expensive the Zen 6 chiplet size has remained largely the same.But the $500 chip wont be more competitive, because of typical Intel segmentation that cuts the amount of bLLC there.
Depends.... the 14900k goes almost toe to toe with the 7800x3d even with almost 1/3 the L3 cache. Also, the 14600k is almost equivalent to the entire Zen 4 non-x3d lineup and the 14700k beats the entire Zen 4 lineup. I feel like it's a wait and see, even with lets say double the normal L3 cache there is a good chance Intel catches AMD but only time will tell. Not enough info is revealed yet to really give any real details of gaming performance...
www.techpowerup.com/review/intel-core-ultra-9-285k/29.html
www.techpowerup.com/review/amd-ryzen-9-9950x/26.html
When the 285K reaches 24t the average clock speed is ~4900 MHz, but the P-Cores are at ~5400 MHz and E-Cores always at 4600 MHz.
For the 9950X when it reaches 32t the average clock speed is ~5300 MHz.
So not sure why you say that E-Cores are 1.1 GHz slower.
Anyway, Nova Lake could have higher E-Core clock speed compared to Arrow Lake but it all has to fit within a reasonable power draw unless Intel pulls another Raptor Lake when going for the crown.
That's why, also taking into account the core advantage/thread parity vs Zen6, I think the strategy with Nova Lake is to keep the clock speed rather tame and rely on core counts to win, keeping the clock speeds "moderate" ensures the chip won't go Super Nova.
Meanwhile Zen6 is rumored to aim for very high clock speeds, that doesn't mean it will be true, however AMD might be able to have a clock speed advantage and keep a similar maximum power draw. Depending on how much higher the clock speed will be in some workloads that don't scale all the way with thread count (still MT but not heavy MT) Zen6 can win.
Right now the 285K draws a bit more than the 9950X in heavy MT. Even though it's on a smaller node and with pretty much equal clock speed (on the P-Cores).
Apparently Nova Lake and Zen6 will be on the same node or a very similar one, some time ago it was rumored that Nova Lake will be N2P and Zen6 N2X but we'll see what exactly will happen.
But if the goal is to have a similar power draw behavior in heavy MT of the 485K vs 10950X compared to the 285K vs 9950X then realistically it seems that Nova Lake has to have somewhat lower clock speed than Zen6.Definitely not impossible. Actually now that I think more about it I would reckon that Zen6 vanilla can match Zen5 X3D due to primarily higher clock speed (and IPC gains) not necessarily due to those increased caches. I would say that if you want to brute force performance past a certain ceiling then clock speed is the likely winning ticket compared to increasing cache. And once that is maxed out relative to power draw targets X3D-ing the new arch pushes things another step further.It was said that there will be a bLLC model with only one compute tile, same core config as the 285K, whether that will competitive enough remains to be seen. Price wise this will definitely be more affordable than the top part, still if the performance is lower than that of the 10800X3D then price has to be lower as well to at least match it in price to performance.
The point is that the performance gaps must be smaller than today, even if AMD still wins, Intel being a close second means a viable alternative, for the right price of course. If they trade blows and a stalemate is practically reached even better for us.
Ideally Intel will price 265K competitively with Zen 6 and does not cut too much of bLLC on it.
Intel also needs to come out and say how many generations will their newest socket support.
A firm statement in this regard could go a long way to convince people to switch to Intel.
www.techpowerup.com/342543/more-evidence-for-avx10-and-apx-support-in-intel-nova-lake-emerge
One possibility is that it'll only be in the P cores, since AVX-512 seems to be an extremely power-hungry and space-consuming feature. Now that Intel and Microsoft have improved scheduling, it wouldn't be a big hurdle to automatically put AVX-512 workloads only on the P cores. But that's just a guess, we'll have to wait and see.
And BLLC is there to cover the outliers that made even 5800X3D 30% fasterAccording to Raichu, Arctic Wolf has feature parity with the P core in client.
Speculation of others is that Arctic Wolf is closer to a zen4C core design with higher IPC
www.phoronix.com/news/Intel-Diamond-Rapids-APX-AVX10
As far as I understand, due to the limitations of the e cores, nova lake will have avx2+many of the instructions of avx512 bit not fp/int 512. The e cores might use 2x 256 units like zen4, who knows.
Also you missed the news that intel is redirecting much of their cpu capacity to AI thus sending high end/workstation cpu prices to the moon. We should expect price increases by the end of the year, right?
www.techpowerup.com/345535/intel-reallocates-pc-production-capacity-to-server-cpus-amid-tight-wafer-supply
Two Nova lake p cores will be as strong as 4e cores allegedly (p cores are clusters of two now with shared l2 like e cores are clusters of four with shared cache) and they do not have ht. We have 16p+32e ≈ 32p cores (+4lpe) without HT VS 24 cores with smt. Intel will have the multithread advantage probably but will not come cheap and the power consumption will be insane. (probably not that hard to cool though because the dies are large)
Zen 6 vs nova lake is going to be interesting because both architectures address the main weaknesses of the previous 4. For zen is the infinity fabric latency and ram speed and for nova is the low p core count and instruction set (maybe?)