GEEKOM Links Four A9 Mega PCs to Power Local AI Clusters

Each unit in the cluster relies on the AMD Ryzen AI Max+ 395, which integrates 16 Zen 5 CPU cores and Radeon 8060S graphics. By utilizing Ubuntu, ROCm, and DwarfStar, the configuration distributes the model across the four machines. The USB4 connection serves as the backbone, removing the necessity for proprietary high-speed switches or bulky server racks typically found in professional environments.

For organizations handling internal policy analysis or source-code review, the platform supports contexts up to 250,000 tokens. Testing shows the cluster achieves 14.61 tokens per second at single concurrency, with a P95 time to first token of 0.42 seconds. Beyond raw speed, the modular nature of the hardware allows companies to scale from a single unit to a four-node deployment as their computational requirements evolve.