AMD has introduced the Ryzen AI Max 400 series, a refreshed version of its Ryzen AI Max 300 “Strix Halo” platform aimed mainly at AI development machines and workstation-style systems.
This is not your normal thin-and-light laptop chip. The Ryzen AI Max idea is basically AMD squeezing a serious CPU, strong integrated graphics, AI acceleration, and a very wide unified memory setup into one package. For Malaysia and SEA, that matters because more creators, students, indie devs, and small studios are trying to run AI workloads locally without jumping straight into expensive desktop GPU rigs.
The big upgrade: up to 192GB unified memory
The headline change is memory. Ryzen AI Max 400 now supports up to 192GB of LPDDR5X memory, up from the 128GB ceiling on the Ryzen AI Max 300 series.
Because this is unified memory, the system and integrated GPU share the same pool. AMD also lets users manually split that memory between normal system use and graphics use. In the most extreme setup, the iGPU can be given up to 160GB of video memory.
That sounds gila on paper, but the use case is pretty clear: AI development, large local models, heavy creative pipelines, and workstation tasks where memory capacity can matter more than just raw gaming FPS. For Malaysian AI students, small software teams, or creators experimenting with local generative tools, this could make compact PCs and dev boxes more practical.
Still a serious CPU and iGPU combo
The chip can come with up to 16 full Zen 5 CPU cores, so this is not some lightweight productivity part. AMD has also nudged the CPU boost clock slightly higher, with the new series reaching up to 5.20GHz compared with 5.10GHz on the previous generation.
On the graphics side, Ryzen AI Max 400 keeps a large integrated GPU based on RDNA 3.5, with up to 40 compute units. AMD has also raised the maximum GPU boost clock to 3.00GHz, compared with 2.90GHz before.
No, this does not automatically mean it replaces every gaming laptop with a discrete GPU. But for compact workstations, mini-PC-style machines, dev platforms, and creator systems, a beefy iGPU with a huge shared memory pool is very interesting. It could also be useful for esports content teams or production setups that need one small machine for editing, encoding, light 3D, and AI tools.
NPU gets a small bump too
AMD says the integrated NPU now delivers 55 TOPS, roughly a 10% improvement over the earlier 50 TOPS-class design. The NPU is there for dedicated AI acceleration, especially for workloads that can run efficiently without hammering the CPU or GPU.
The platform still uses a 4-channel, 256-bit wide LPDDR5X unified memory interface. That wide memory setup is a major part of why Strix Halo has been interesting in the first place.
Business first, consumer versions later
AMD is launching three Ryzen AI Max 400 models, and all three include AMD PRO features. That puts them in the business and enterprise lane, similar to how Intel positions vPro for manageability and security.
For normal consumers, AMD says consumer versions are planned later this year. That is the part local buyers should watch. If brands bring these into Malaysia in mini PCs, creator laptops, or compact workstation boxes, pricing will decide whether this becomes a niche AI dev flex or something more accessible.
For now, Ryzen AI Max 400 looks less like a gaming-first launch and more like AMD sharpening Strix Halo for the AI workstation crowd. Bigger memory, slightly faster clocks, and a stronger NPU may not sound flashy, but for local AI development and creative workloads, this is exactly the kind of spec bump that can make a small machine feel much less limited.
Source: TechPowerUp