AMD and Intel are developing new CPUs with increased L3 cache, potentially targeting local large language model execution rather than solely gaming performance. This development follows AMD’s introduction of 3D V-Cache technology in its Ryzen 5000 series, which provided a gaming advantage.
Intel’s upcoming response to AMD’s 3D V-Cache technology is expected with its Nova Lake CPU designs. A recent leak indicated Intel is developing a CPU featuring two “Big Last Level Cache” (BLLC) chips, functioning as Intel’s equivalent of 3D V-Cache. Shortly after this Intel revelation, an AMD leak disclosed the company is also working on a dual CCD V-Cache CPU, suggesting a strategic move by AMD to pre-empt Intel’s cache-related announcements.
The advanced AMD CPU featuring dual CCD V-Cache could launch as early as 2024. In contrast, Intel’s BLLC alternative is not projected for release until 2026, coinciding with the anticipated debut of its Nova Lake CPU line. This timeline places AMD’s potential launch two years ahead of Intel’s comparable offering.
In January 2025, AMD stated that developing such a solution was not economically viable and is a highly specialized niche. Now I can say that it does make sense, because accelerated inference of large language models provides a boost.
— 1usmus 🇺🇦 (@1usmus) August 4, 2025
Despite potential benefits for gaming, the utility of these high-cache CPUs for this application may be limited. AMD previously indicated that incorporating more than one V-Cache chip on its processors would not be cost-effective and would not substantially enhance gaming performance. This assessment is based on the premise that most contemporary games do not fully utilize the eight cores present on a single Ryzen chiplet in AMD’s top-tier processors.
While certain games might leverage more than eight cores and could potentially benefit from additional V-Cache across multiple core collections, the primary design intent for these chips may be centered on accelerating artificial intelligence performance. 1usmus, a notable AMD overclocking tool developer, stated that dual V-Cache CPUs are relevant in the current AI landscape due to their capacity to accelerate inference processes within large language models.
Additionally, concerns exist regarding asynchronous cache across cores in the Windows operating environment, which could lead to performance degradation over time. Reports suggest that Microsoft may have limited options to address this issue. A dual V-Cache configuration could offer a solution for both AMD and Intel by simplifying future designs while simultaneously enhancing AI acceleration capabilities.
