Intel reportedly plans “Unified Core” architecture to succeed Razor Lake

Intel’s reportedly planning to ditch P-cores and E-cores, opting for a new “Unified Core” design

Rumour has it that Intel is planning a drastic change for its future CPU designs, with @Silicon_Fly reporting that Razer Lake will be Intel’s last CPU generation to use P-Cores and E-cores. Razer Lake is Intel’s planned successor to next year’s Nova Lake CPUs. After Razer Lake, Intel reportedly plans to launch Titan Lake, which will reportedly feature a new “Unified Core” architecture. These CPUs are due to be released in 2028.

With this new architecture, Intel will remove the downsides of having a multi-architecture CPU design. When CPUs have P-cores and E-cores, balancing work between these cores can be a challenge. Are tasks that require the highest levels of single-threaded performance always on the fastest cores? Is the overhead of core management dropping performance? Are apps multi-threaded enough to take advantage of HUGE E-core counts?

Having a unified core architecture has its perks

For the most part, AMD has opted to use only a single CPU core architecture in its processors. While some of AMD’s CPUs feature both Zen and Zen c cores, these CPUs account for a minority of AMD’s designs. It is also worth noting that these configurations are not found in any high-end desktop CPUs or server CPUs. Furthermore, AMD’s Zen and Zen c series cores are largely similar, with the differences between them being core density, clock speeds, and cache allocation. They are not fundamentally different core designs, like Intel’s P-cores and E-cores. Both AMD’s Zen and Zen c cores support the same instruction sets and core functionalities. The same isn’t true for Intel’s P-cores and E-cores. This results in problems, such as Alder Lake/Raptor Lake’s lack of proper AVX-512 support.

It is claimed that Intel’s Unified Core design will be based on the company’s E-core designs. Although Intel’s Unified Core design will see Intel’s E-cores increase in size, as they will replace Intel’s P-cores, they are expected to deliver idealised performance per area, per watt, and raw performance characteristics.

Intel Unified Core:

Most of the info is already available. But a new leak confirms most of it and also sheds more light.

It clearly says that Razer Lake's Griffin Cove is the *LAST* Intel P Core. And after that, the P core team is going bye-bye.

And in 2028, Titan Lake is all… pic.twitter.com/nny4xYYXMc

— SiliconFly (@Silicon_Fly) July 14, 2025

Is Intel’s “Unified Core” concept a good idea?

There’s no reason why Intel can’t create a CPU core that’s both performant and efficient. While Intel’s P-cores are designed to deliver higher levels of performance, E-cores do not guarantee higher levels of power efficiency. A big core can be power-efficient. Honestly, the biggest perk of Intel’s E-core designs is their spatial efficiency, not their power efficiency. E-cores take up much less space than P-cores, so much so that a single Raptor Lake P-core takes up roughly the same amount of space as four E-cores. Four E-cores can deliver more multi-threaded performance than a single P-core. That factor alone is a big reason why Intel has invested in E-cores.

If Intel can create a “Unified Core” that is compact and performant, Intel will be able to deliver incredibly high levels of single-threaded and multi-threaded performance. It will also eliminate the complexity of migrating most workloads between CPU core architectures. Core allocation can be based on factors like clock speeds alone, instead of considering architecture, clock speed, and power.

While Intel is reportedly creating a “Unified Core” architecture, Intel may follow the AMD Zen route with multiple core designs using the same core architecture. There could be Intel Unified Standard and Dense core designs. This can give Intel the perks of P-cores and E-cores without needing to invest in two utterly different core designs.

You can join the discussion on Intel’s rumoured “Unified Core” on the OC3D Forums.