Saturday, August 29th 2026
Intel Confirms LGA1954 as the Socket for "Nova Lake-S" Processors Through Internal Tools
The upcoming Nova Lake-S lineup of Intel desktop CPUs has been the subject of strong rumors for a while, and one of them concerned the socket itself. Intel is moving away from LGA1851 and will have a brand-new socket for its upcoming Core Ultra 400 series processors, namely LGA1954. We already know about LGA1954 as its records appeared in shipping manifests months ago, and we even know what the socket looks like. Nova Lake-S will debut on this new socket, and Intel has now officially confirmed this, though not in the usual way.
Intel has just published an internal test tool on its website that connects the LGA1954 socket to Nova Lake-S processors. This listing is for validation hardware and provides some engineering and technical enablement equipment. Nevertheless, the important part is that this is the first time Intel itself has mentioned Nova Lake-S alongside LGA1954 in any sort of official capacity. This is confirmation that Nova Lake-S will indeed launch alongside the LGA1954 socket and Intel 900-series motherboards. Nova Lake is currently expected to launch in Q1 2027.
Source:
Harukaze5719 on X
Intel has just published an internal test tool on its website that connects the LGA1954 socket to Nova Lake-S processors. This listing is for validation hardware and provides some engineering and technical enablement equipment. Nevertheless, the important part is that this is the first time Intel itself has mentioned Nova Lake-S alongside LGA1954 in any sort of official capacity. This is confirmation that Nova Lake-S will indeed launch alongside the LGA1954 socket and Intel 900-series motherboards. Nova Lake is currently expected to launch in Q1 2027.



87 Comments on Intel Confirms LGA1954 as the Socket for "Nova Lake-S" Processors Through Internal Tools
Socket longevity is overrated.
But for each person this can be different.
There are lots of reasons why someone may change their boards while the one they have technically support the CPU they upgrade to. Motherboards don't release at the same time and brands may release a "v2" later in the cycle.
While upgradability is a factor, it's overrated because you don't know exactly what's coming in the incoming years and when precisely it'll come, and for how long the CPU manufacturer will keep its hardware compatible. In the case of AMD, it has just been about bullying them in order to keep Zen3 on AM4, so we had an estimation of the date because people knew this would be the last generation on AM4.
With B650 it varies based on the model, either Gen4 or Gen5 GPU slot.
While we're firmly out of modern sockets in the total lifespan discussion, AMD's Slot A gets a boost toward the podium even though it technically housed four variants of K7, it did so in 11 months (and 14 days).
The main reason to replace a motherboard is the CPU. A new architecture can bring double the cores, or 20% performance increase from that new architecture. Someone who replaces a motherboard because of a new USB standard, well, probably doesn't know why those PCIe slots are on the motherboard for.Windows got much better on this the last many years, basically from Windows 10 and latter, still replacing the motherboard isn't 100% warranted that you wouldn't end up troubleshooting problems that weren't there with the old setup. And if the person using the PC isn't someone with tech knowledge, what will be the cost of upgrading simply the CPU, compared to a motherboard? What will the technician charge in the first case and what in the second one? Let's not limit conversation to ourselves alone.
There would be a point if AM5 offered better IO, which it doesn't, so this advantage is at best exaggerated. And a such advantage would be for buyers in the beginning of the platform, not buyers considering LGA1954 vs. AM5 at the tail end, where the platform becomes more of a limitation than an advantage.
Buy whichever performs the best for your use case, not a theoretical edge case where you upgrade the least likely part to be upgraded, and keep the most obsolete part of the system. :rolleyes:For higher throughput and lower effective latency.
You remember this one? You posted there.
A ticket to the concert yes, but not a front row experience. People have to plan better.
1954 needs to have long term support. Most indications are, it will. Flagship gaming performance will go a long way, but not as far as assurances it's just the beginning for the socket.
Your windows license gets de-activated but it can be re-activated for free so there's that.
My currently Win11 install started life as a Win7 install on my very first SSD when I built an i5-3570K based system. It has since been upgraded to 10 and 11, cloned across at least 5 new SSDs, and called at least a dozen different motherboards home across both AMD and Intel over the years. Modern Windows just doesn’t care.
I think they should just change to the thermalright type contact frame, consumer installation is not much harder than a M.2, and has even without excess pressure installation alongside a better looking board, I would imagine it could be made clip on or something for even easier installation.
These existing latches must be cheaper as is the only way I can think of them to keep using them.
Admittedly the IHS is still not perfectly flat but way better than the old one.
Unfortunately not every LGA1851 board has the RL-ILM, some cheaper B860 ones have the old one. You can actually tell from pictures, even the official ones if the resolution is high enough, if not then from reviews.
www.techpowerup.com/review/intel-core-ultra-7-270k-plus/26.html
Fortunately even if the IHS is not perfectly flat using the RL-ILM for chips such as 245K and 250K a contact frame is not needed, the temps are too low to worry about.
For the 265K it's 50-50 in my opinion, a contact frame could improve things a bit, but the temps are very reasonable, so it makes more sense for an air cooler setup, an air cooler being less capable than a 360 AIO.
For the full chips however, the 270K and 285K a contact frame would yield somewhat larger improvements which would matter more relative to the higher temps.
So I wouldn't say mandatory but recommended.
Regarding the new two lever ILM I'm guessing that will achieve more IHS flatness due to the symmetry, but something else which could also help is the shape of the IHS. I'm expecting LGA1954 to be obviously larger, but closer to a square than LGA1851 which was narrower but longer than LGA1700.
As such I don't think a contact frame will be really needed for LGA1954.