AMD is pushing its AI PC chips into properly workstation-level territory with the new Ryzen AI MAX 400 family, also known as Gorgon Halo.
The big headline is not just “AI PC” branding — we have plenty of that already. The real upgrade is memory. AMD says the Ryzen AI MAX 400 SoCs can support up to 192GB of unified memory, which gives these chips enough headroom to run very large local AI models, including 300B+ parameter LLMs, on an x86 client processor.
For Malaysian and SEA developers, creators, students, and small studios, that matters because local AI is becoming more than a flex. Running models on your own machine can help with privacy, lower cloud dependency, and faster iteration when you are building tools, testing agents, editing media, or experimenting with AI workflows without constantly paying API bills.
What is inside Ryzen AI MAX 400?
The Ryzen AI MAX PRO 400 and Ryzen AI MAX 400 chips sit under AMD’s Gorgon Halo family. The architecture is familiar if you followed Ryzen AI MAX 300: Zen 5 CPU cores, RDNA 3.5 graphics, and an XDNA 2 NPU.
So this is not a total architecture reset. Instead, AMD is tuning the platform with higher clocks and much bigger memory support. That memory jump is the spicy part, bro.
AMD is positioning these chips for:
- Large local AI models, including 300B+ parameter LLMs
- Multiple local AI agents running at the same time
- Creative and professional workloads
- Design, rendering, simulation, and engineering applications
- Systems that combine compute, graphics, and AI acceleration in one platform
Another key detail: users can allocate up to 160GB of VRAM to the GPU. That is a big step up from the 112GB allocation possible on current 128GB setups, and it gives the integrated Radeon graphics far more breathing room for AI and professional workloads.
Three chips at launch
According to the source details, the Ryzen AI MAX 400 family will start with three chips:
- Ryzen AI MAX+ PRO 495
- Ryzen AI MAX+ PRO 490
- Ryzen AI MAX+ PRO 485
These match the core configurations of existing Ryzen AI MAX 300 models, but AMD is giving them higher CPU and GPU clocks.
The flagship Ryzen AI MAX+ 495 comes with 16 Zen 5 CPU cores and a Radeon 8065S iGPU with 40 compute units. CPU clocks are up by 100MHz, landing at 3.1GHz base and 5.2GHz boost. The GPU also gets a 100MHz bump, reaching 3.0GHz. The NPU is rated at 55 TOPS.
Power-wise, the chip has a 55W base TDP, configurable down to 45W and up to 120W depending on the system design.
Why Malaysia should care
This will not be your average budget gaming laptop chip. No Malaysia or RM pricing has been announced yet, and with ASUS, HP, and Lenovo listed as launch partners, expect these to first show up in premium AI laptops, mobile workstations, and creator-focused machines.
For Malaysia’s dev scene, AI startups, university labs, content creators, and even indie game teams, the appeal is clear: more serious local AI without needing a full desktop workstation or cloud GPU rental all the time. If OEMs bring these systems into Malaysia at reasonable RM pricing, this could be a genuinely useful option for people doing AI-assisted coding, video production, 3D work, or local model testing.
Systems using Ryzen AI MAX PRO 400 processors are expected from major OEMs including ASUS, HP, and Lenovo starting Q3 2026. AMD’s standard Ryzen AI 400 “Gorgon” chips are also expected soon, aimed at competing with Intel’s Core Ultra Series 3 “Panther Lake” lineup.
Source: Wccftech Gaming