Framework is preparing to open pre-orders for a new high-end Framework Desktop powered by AMD’s Ryzen AI MAX+ 495, and the standout feature is its massive 192 GB of unified memory. Pre-orders are scheduled to begin this Wednesday morning Pacific Time, positioning the compact desktop as one of the most interesting upcoming systems for AI developers, workstation users, and enthusiasts working with large language models.
The new Framework Desktop is built around AMD’s Ryzen AI MAX+ Pro 495, a powerful chip aimed at local AI workloads, content creation, development, and demanding professional use. It features 16 Zen 5 CPU cores and reaches clock speeds of up to 5.2 GHz, giving it a notable performance edge over the earlier Ryzen AI MAX+ Pro 395. On the graphics side, the chip includes Radeon 8065S integrated graphics with 40 RDNA 3.5 compute units running at up to 3.0 GHz.
The biggest upgrade, however, is memory. Framework’s new configuration includes 192 GB of LPDDR5X unified memory running at 8533 MT/s, delivering up to 273 GB/s of bandwidth. Compared with previous Ryzen AI MAX+ 395 systems, which topped out at 128 GB and 256 GB/s of memory bandwidth, the new model offers a 50% increase in capacity and a modest boost in bandwidth.
That memory capacity is especially important for AI users. Local AI models and large language models can demand huge amounts of memory, and Framework says the system is optimized for running advanced LLM workloads. The company has highlighted the ability to run DeepSeek V4-Flash at Q8 on a single desktop while still leaving room for additional context length.
Framework is also making useful hardware changes to the desktop design. The PCIe x4 slot is now open-ended, allowing users to install physically larger PCIe cards directly into the system. This gives the machine more flexibility for networking, storage, expansion cards, and specialized workstation configurations.
To demonstrate the potential of the open-ended PCIe slot, Framework showed a concept setup using two 50 GbE network interface cards. With RDMA over Ethernet support, such a setup can enable extremely low-latency communication between systems. In practical terms, this could allow two Framework Desktop units to work together and effectively pool 384 GB of memory for AI workloads using tensor parallelism.
Framework also plans to offer a pre-built configuration with Fedora installed, giving Linux users a ready-to-go option out of the box. This could make the system more appealing to developers, researchers, and AI professionals who prefer a Linux-based environment for machine learning, coding, and server-style workflows.
Pricing has not been officially confirmed for the 192 GB model, but Framework has warned that rising LPDDR5X memory costs will make this configuration significantly more expensive than the existing 128 GB version. The current Ryzen AI MAX+ 395 Framework Desktop starts at around $3,500, while other high-end Ryzen AI MAX systems are often priced in a similar range. Early listings for comparable 192 GB systems have appeared above $6,000, so the new Framework model is expected to land in a premium price bracket.
Framework is also advising buyers to consider the 32 GB, 64 GB, or 128 GB configurations if budget is a concern. Still, for users who need as much unified memory as possible in a compact AI-focused desktop, the 192 GB Ryzen AI MAX+ 495 model could be a compelling option.
With a powerful Zen 5 processor, upgraded Radeon graphics, 192 GB of high-speed unified memory, Linux support, and flexible expansion, the new Framework Desktop is shaping up to be one of the most capable compact AI PCs on the market. For developers working with large language models, researchers building local AI systems, and professionals who need workstation-class performance in a small form factor, this upcoming pre-order could be worth watching closely.
