Leo · 2026-11
woCyber’s X4 is a published 4-bay private cloud appliance with local AI: Intel N100, 8 GB DDR4, 64 GB eMMC boot storage, four tool-free SATA trays, a 120 TB raw capacity ceiling and 2 × 2.5GbE. It is the base configuration woCyber documents and reconfigures for OEM buyers looking for an Intel N100 2 × 2.5GbE NAS ODM.
For a procurement team comparing private-label platforms, the published parameter list is the starting point rather than a promise of measured performance. RAID 0/1/5/6/10, NVMe cache bays and 10GbE port mixes are listed as OEM-reconfigurable options; every other change is handled as an engineering feasibility request.
The published specification lists an Intel N100 processor (4 cores, up to 3.4 GHz), Intel UHD Graphics with hardware 4K transcoding, 8 GB DDR4 memory and 64 GB eMMC boot storage. Storage is four tool-free 2.5-inch or 3.5-inch SATA HDD or SSD trays with a 120 TB raw ceiling (30 TB × 4). Networking is two independent 2.5GbE RJ45 ports, plus USB 3.2, Type-C and HDMI output at 4K@60 Hz. The chassis is aluminum over a steel mid-frame with active multi-fan thermal control. Read the X4 full specification for the complete parameter list.
Two independent 2.5GbE ports support 802.3ad link aggregation and failover, giving up to 5 Gbps of aggregated interface bandwidth for concurrent multi-user workloads. Those are interface rates, not measured transfer speeds: sustained throughput depends on drives, RAID level, file sizes and client hardware. The X4 product documentation also lists 10GbE port mixes as an OEM-reconfigurable option, so a request for an SFP+ or 10GbE interface is handled as an engineering feasibility question rather than assumed as a standard SKU.
Keeping data on-premise means the appliance, not a cloud region, becomes the trust boundary. On the X4 the local compute path is the Intel N100 with Intel UHD Graphics for hardware transcoding; local RAG and lightweight inference workloads are commonly offloaded to that iGPU or to an external accelerator node. woCyber publishes the hardware platform and its interfaces, not a benchmark of any particular model or framework. An integrator should validate the intended inference stack on a sample unit before committing to a rollout, and should state the required encryption, key management and remote-access behaviour in the project brief.
The published manufacturing process places SMT lines, assembly, a temperature-controlled 72-hour burn-in room and outgoing inspection under one quality system, with batch traceability against serial numbers and 100% OQC before packing. For a remote or unattended deployment, validation before shipping is what keeps a storage failure from turning into a field service visit. The site also publishes a 3-year warranty, 500K+ units shipped, 20,000+ units of monthly capacity and 15+ years in the industry.
The published MOQ for reconfiguring an existing platform is 100 units per SKU. A new PCB respin or a new enclosure tooling program starts at 1,000 units. The four documented customisation levels are logo and brand UI, enclosure finish and packaging, new PCB respin, and new enclosure tooling. Volume pricing and tiered prices are not published; volume supply is handled on request. These are order quantities, not delivery lead times.
Open the DIY Configurator on the platform matrix, select a base platform, compute and memory, network interface, AI acceleration and storage bundle, and customisation scope. The tool generates a reproducible Config ID from those selections. Submit it with the inquiry and we reply with an engineering feasibility assessment within 48 hours. Related reading: the White-Label NVR Surveillance Storage Architecture Brief and the Architecture Briefs library.
Browse all platforms →An off-grid security camera with local storage records onto hardware at the site, powered by the site's own supply, with no cloud…
November 20, 2026A NAS backup strategy for business rests on three pillars in 2026: a copy that is separate from the primary array, a…
November 20, 2026A security camera system without monthly fees records to storage you own, on hardware that processes video locally, instead of uploading clips…
Ask our export team for the OEM/ODM cooperation guide and a written quotation.