QAI-h1290FX, QNAP's Solution to Bring AI to the Edge

QNAP has made available the QAI-h1290FX, a high-end NAS designed to operate as a storage server ready to run AI workloads at the edge. It supports NVIDIA RTX PRO Blackwell GPUs (not included in the standard configuration), which allow it to execute inference workloads (for example, to run a local AI chatbot) and RAG research, as well as more traditional NAS tasks like virtualization.

QNAP Brings AI to the Edge with QAI-h1290FX

Although QNAP indicates that the QAI-h1290FX is a server designed for storage, its capabilities go far beyond that. Based on the AMD EPYC 7302P 16-core CPU and 128 GB of RAM (expandable up to 1 TB), it features 12 SATA and NVMe drive bays. Connectivity is managed by two 25 GbE ports and two 2.5 GbE ports. If that’s not enough, a network adapter can be installed in one of the available PCI slots, adding 100 GbE ports.

The most interesting aspect is the possibility to install an NVIDIA GPU, specifically an NVIDIA RTX Pro 4500 Blackwell with 32 GB of RAM or, for heavier tasks, an NVIDIA RTX Pro 6000 Blackwell Max-Q equipped with 96 GB of RAM. The former is ideal for locally running medium-sized LLMs, that is, up to a maximum of 30 billion parameters, while the latter is suited for more intensive AI workloads, such as managing LLMs with over 70 billion parameters. The operating system is QuTS Hero, which relies on the ZFS filesystem.

Among the typical applications for this type of machine are AI chatbots, which will have the advantage of being run locally, ensuring complete privacy. For more in-depth analyses, QAI-h1290FX can handle RAG systems to verify the correctness of AI outputs. It can also be an excellent companion for creatives: the NVIDIA RTX PRO GPU allows it to generate images using models like Stable Diffusion.