GEEKOM has assembled a four-node AI cluster using its A9 Mega Mini PCs, running DeepSeek V4 Flash for local inference. This setup connects four A9 Mega systems via USB4, providing a distributed platform for AI workloads without the need for a traditional data-center server.
Each node in the cluster is powered by the AMD Ryzen AI Max+ 395, featuring 16 Zen 5 CPU cores, Radeon 8060S graphics, and unified memory. The system runs on Ubuntu with ROCm and DwarfStar to distribute an optimized DeepSeek V4 Flash model across the four systems. This configuration allows applications and AI agents to connect through an OpenAI-compatible API.
One of the key advantages of this setup is its ability to keep AI processing and sensitive information within the organization’s own infrastructure. Prompts, documents, source code, credentials, and intermediate results remain local, which is crucial for organizations handling sensitive data. This makes it ideal for private knowledge assistants that work with internal policies, manuals, contracts, and reports. Other supported workloads include document review, source-code analysis, local retrieval-augmented generation (RAG), multi-source research, and controlled workflow automation.
The system can handle up to 250K tokens for agent workloads, making it suitable for long documents, large codebases, and complex jobs. In testing, the four-node system delivered approximately 14.61 tokens per second at single concurrency, with a P95 time to first token of around 0.42 seconds in the 32- and 128-token tests. GEEKOM emphasizes that this configuration is designed to provide additional capacity and acceleration for long prompts rather than simply increasing text-generation speed.
One of the flexible aspects of this setup is that organizations do not need to deploy all four A9 Mega systems at once. They can start with one or two systems and add additional nodes as their AI workloads grow. Each A9 Mega can operate independently or contribute to distributed inference, making it a scalable solution for businesses, laboratories, classrooms, and edge environments.
For those shopping for a mini PC, the GEEKOM A9 Mega AI cluster offers a compelling option for local AI processing. The use of USB4 for connectivity ensures high-speed data transfer without the need for a proprietary high-speed switch or server rack. This makes it a cost-effective and flexible solution for organizations looking to deploy local AI without moving directly to a traditional server-based infrastructure.
For more details, you can check out the original article here.
