The Tiny Powerhouse: Gaming and Local AI

Thu Feb 27 2025
A well-known brand for its modular, fixable laptops has just released its first desktop computer. This compact desktop PC is packed with impressive features. It is designed to handle both gaming and running large language models locally. The standout feature of this desktop is its internal components. It is one of the first to use AMD's latest Strix Halo architecture, known as the Ryzen AI Max processors. This all-in-one processing unit promises high performance. The Framework Desktop is incredibly small, measuring just 4. 5L. It is built around a mini-ITX mainboard, making it smaller than popular gaming consoles like the PlayStation 5 or Xbox Series X. The front panel is customizable with 21 interchangeable plastic tiles, allowing users to choose their own colors and patterns. The desktop comes with a variety of expansion cards, including USB-C, USB-A ports, a headphone jack, an SD card reader, and a storage expansion card. These cards are placed at the bottom of the device, offering flexibility and convenience. Inside, the desktop is quite straightforward. It includes a mainboard with AMD's accelerated processing unit, a fan, a heat sink, a power supply, and two M. 2 2280 NVMe SSD slots for storage. The AMD Strix Halo APU is soldered to the mainboard, with two configurations available: the AMD Ryzen AI Max 385 and the AMD Ryzen AI Max+ 395. The top configuration boasts 16 CPU cores, 40 graphics cores, and 80MB of cache, while the entry-level configuration has 8 CPU cores, 32 graphics cores, and 40MB of cache. One notable design choice is the soldered-in RAM, which ranges from 32GB to 128GB. This means users cannot upgrade the RAM in the future. The decision to solder the RAM was made to achieve the high memory bandwidth required by the Ryzen AI Max processors. Despite this limitation, the large amount of unified memory opens up possibilities for running large language models locally. The desktop supports various open-source tools for local AI workloads, such as Llama 3. 3 70B, Ollama, llama. cpp, and other models from Mistral, Nous, Hermes, and DeepSeek. Additionally, the mainboard can be purchased without a case, allowing for custom setups like a mini-rack with four mainboards running in parallel for AI testing. The base model of the Framework Desktop starts at $1, 099, while the top-end version costs $1, 999. It supports both Windows and popular Linux distributions like Ubuntu, Fedora, and Bazzite. Pre-orders are now open, but shipments will start in early Q3 2025. This desktop is a game-changer for those looking for a compact, powerful machine. It offers a unique blend of gaming performance and AI capabilities, making it a versatile choice for tech enthusiasts. However, the soldered-in RAM may be a deal-breaker for some users who prefer the flexibility of upgradable components. The customizable design and modular expansion cards add to its appeal, allowing users to tailor the desktop to their specific needs. Overall, the Framework Desktop is a promising addition to the market, offering a lot of power in a small package.
https://localnews.ai/article/the-tiny-powerhouse-gaming-and-local-ai-1235910d

questions

    How does the performance of the Framework Desktop in running large language models compare to other specialized AI hardware solutions on the market?
    How does the lack of RAM upgradability affect the long-term value and usability of the Framework Desktop for users with evolving needs?
    What are the potential environmental impacts of using soldered-in components versus modular, upgradable parts in the Framework Desktop?

actions