Framework Desktop Now Available for Pre-order: AMD Ryzen AI Max 400 Series with Up to 192GB RAM

Framework Desktop Gets AMD Ryzen AI Max 400 Series

Modular hardware maker Framework has opened pre-orders for its new Framework Desktop, headlined by the AMD Ryzen™ AI Max 400 series processor and support for up to 192GB of memory. For users focused on local AI inference and high-memory workloads, this combination is a compelling one.

The pre-order lineup includes the standard Framework Desktop and a Framework Desktop DIY Edition for enthusiasts — both offering the 192GB memory configuration. Framework has long been known for its repairable, upgradeable modular design philosophy, and bringing AMD's latest AI-accelerated platform to a desktop form factor continues its focus on power users.

Why 192GB of RAM Matters

For users running large language models (LLMs) locally or performing AI inference, memory capacity is often the decisive bottleneck. The AMD Ryzen AI Max series uses a unified memory architecture (UMA) that allows the integrated GPU and CPU to share a large, common memory pool. In theory, a 192GB configuration can handle models with a considerable number of parameters — without the need to stack expensive discrete GPUs for VRAM.

This type of compact machine with high memory capacity and integrated AI acceleration is well-timed for the growing demand for "local AI." Rather than following the traditional high-end gaming PC route, this Framework product is clearly aimed at AI developers, researchers, and tech enthusiasts who want to run models locally.

The AMD Ryzen AI Max 400 series (codenamed "Strix Halo") and its Unified Memory Architecture (UMA) are central to understanding its AI inference advantages. Unlike traditional PC architectures, this platform integrates the CPU, an integrated GPU (up to 40 compute units), and an NPU (Neural Processing Unit) on a single chip, all sharing the same physical memory pool through an extremely high-bandwidth internal interconnect. This approach mirrors Apple's M-series chip design, but runs on the x86 ecosystem — meaning existing Linux/Windows AI toolchains work natively.

For LLM inference, memory bandwidth and capacity are the core bottlenecks: a 7-billion-parameter model loaded at FP16 precision requires roughly 14GB of VRAM, while a 70-billion-parameter model needs upwards of 70GB. On a traditional PC, achieving the latter requires stacking multiple flagship discrete GPUs — an extremely costly solution. A 192GB unified memory pool can theoretically support models with hundreds of billions of parameters running at 4-bit quantization, which is the core value proposition of this configuration for local AI users.

What the Modular and DIY Edition Mean

The DIY Edition continues Framework's core philosophy — letting users select and assemble their own components, lowering the barrier to future upgrades and repairs. This approach has earned a strong reputation in the laptop space, and bringing it to a desktop platform means users can configure memory, storage, and other components to their specific needs.

For those who want a long-term machine they can incrementally upgrade, this repairability is a clear advantage over all-in-one sealed products. That said, it's worth noting that the information currently available is limited to the pre-order announcement and core specs (AMD Ryzen AI Max 400 series, 192GB memory). More detailed pricing, specific model tiers, shipping timelines, and real-world performance benchmarks are still pending from Framework.

Framework's modular system is engineered differently from typical user-serviceable designs. Its motherboards use a standardized interface specification with publicly available documentation, allowing users and third-party developers to design expansion accessories around it. On the laptop side, Framework has already introduced a swappable "Expansion Card" system that lets users customize their port selection across USB, HDMI, SD card, and more. This desktop product carries that philosophy forward, though the exact extent of modularization — for example, whether the motherboard can be purchased and replaced independently — will need to be confirmed once the full specs are released. Notably, Framework has cultivated an open-source hardware community and has historically offered new processor upgrade kits for older motherboards — a post-purchase upgrade path that sets it apart from traditional OEM brands.

Wrap-up

The Framework Desktop with AMD Ryzen AI Max 400 series and up to 192GB of memory is a compact machine built for local AI and high-memory workloads, with the brand's signature modular and DIY ethos intact. It is currently in the pre-order phase, and interested users are advised to wait for more complete specifications and independent reviews before making a purchasing decision.