Intel Nova Lake-S rumored to get bLLC across four unlocked SKUs

Insiders say Intel is planning four unlocked 'Core Ultra 400' SKUs with bLLC (Big Last Level Cache) enhancements for Nova Lake-S, including a claim of 144 MB of bLLC on a single tile.

Two industry insiders have pushed new details about Intel’s unannounced Nova Lake-S desktop family. Jaykihn recently claimed an upcoming Nova Lake-S SKU would carry 144 MB of bLLC (Big Last Level Cache), a supplemental cache pool that could bolster Intel’s competitiveness in the gaming desktop market where AMD has had success with 3D V-Cache. The 144 MB figure surfaced as a notable single-tile boost that analysts say could change performance expectations for certain workloads.

Following that post, kopite7kimi provided additional claims that Intel plans four individual unlocked SKUs as part of the late 2026 Core Ultra 400 CPU family that will feature bLLC implementations. Experts cited in the thread expect the extra cache to be added to the Nova Lake-S design’s compute tile. The discussion also differentiated between expected single-tile configurations and possible dual-tile options, indicating Intel may ship multiple die configurations under the same family to target different segments of the desktop market.

Conversation around die configurations referenced a previously reported 52-core engineering sample that used a dual-tile layout described as two 8+16 tiles plus 4 low power cores. Sources now propose a variant that could implement 144 MB of Big Last Level Cache on each tile, producing a combined 288 MB bLLC in-package. Kopite7kimi cautioned that the information is not final and added a platform-level caveat, writing ‘(I am) not 100% sure yet. The power supply design of the new platform will be a challenge for motherboard manufacturers.’ Another dual-tile possibility discussed would use two 8+12 tiles. Taken together, the tips sketch a lineup of multiple unlocked Nova Lake-S SKUs with varying core counts and significant bLLC provisions, though details remain unconfirmed.

52

Impact Score

Artificial Intelligence speeds quantum encryption threat timeline

Research from Google and Oratomic suggests quantum computers capable of breaking core internet encryption may arrive sooner than expected. Artificial Intelligence played a key role in improving one of the new algorithms, raising fresh urgency around post-quantum security.

New methods aim to improve Large Language Model reasoning

A new study on arXiv outlines algorithmic techniques designed to strengthen Large Language Model reasoning and reduce hallucinations. The work reports better logical consistency and stronger performance on mathematical and coding benchmarks.

Nvidia acquisition of SchedMD raises Slurm neutrality concerns

Nvidia’s purchase of SchedMD has given it control of Slurm, an open-source scheduler that sits at the center of many supercomputing and large-model training systems. Researchers and engineers are watching for signs that support could tilt toward Nvidia hardware over AMD and Intel alternatives.

Mustafa Suleyman says Artificial Intelligence compute growth is still accelerating

Mustafa Suleyman argues that Artificial Intelligence development is being propelled by simultaneous advances in chips, memory, networking, and software efficiency rather than nearing a hard limit. He contends that rising compute capacity and falling deployment costs will push systems beyond chatbots toward more capable agents.

China and the US are leading different Artificial Intelligence races

The US leads in large language models and advanced chips, while China has built a major advantage in robotics and humanoid manufacturing. That balance is shifting as Chinese developers narrow the gap in model performance and both countries push to combine software and machines.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.