HP Inc BN5L2UT#ABA HP Z2 Small Form Factor G1i Workstation
The HP Z2 SFF G1i (BN5L2UT#ABA) is a compact professional workstation built around Intel's 15th-generation Core Ultra 5 245K processor — 14 cores at 3.60 GHz base with boost headroom to 5.20 GHz — paired with 32GB of system memory, a 1TB M.2 PCIe NVMe SSD, and an NVIDIA Quadro RTX A100 GPU. The SmartBuy configuration targets commercial deployments where floor space is constrained but compute requirements are not. For a deep dive into the broader HP Inc workstation line, see the full catalog.
Overview
The Z2 SFF G1i sits at the intersection of ISV-grade compute and a footprint small enough for a standard desktop slot. The 14-core Core Ultra 5 245K handles compute-intensive workloads — surveillance VMS processing, CAD, or AI-inference pipelines — without the rack footprint of a tower. Intel AI Boost NPU delivers up to 13 TOPS of dedicated inferencing capacity, offloading AI tasks from the CPU and GPU to reduce latency on real-time analytics. This makes the Z2 SFF G1i a practical choice in environments running on-premise network video recorder software or edge AI workloads that need sustained throughput rather than burst compute.
The SFF chassis measures 19.00 × 16.00 × 8.00 in (L × H × W) at 13.98 lb — light enough to relocate without a dolly, dimensioned to fit in under-desk or credenza deployments alongside standard peripherals. It pairs well with PoE network switches when anchoring a surveillance head-end or AV control station.
Key Features
- Intel Core Ultra 5 245K — 14 cores, up to 5.20 GHz: The 245K's hybrid architecture separates performance and efficiency cores, letting the OS schedule background tasks (indexing, updates, agent services) on E-cores while VMS decoding or render threads run on P-cores. The 5.20 GHz boost ceiling keeps single-threaded latency low for GUI-bound VMS clients and configuration tools.
- Intel AI Boost NPU (up to 13 TOPS): Dedicated neural-processing hardware in the 245K SoC accelerates AI inferencing independently of the CPU and GPU. In a surveillance head-end running analytics-heavy VMS software, this offload reduces CPU utilization meaningfully on streams using object detection or classification at the edge.
- 32GB System Memory: 32GB is the practical floor for running a multi-channel VMS server, a hypervisor hosting two VMs, or a CAD/simulation workstation. It avoids the memory-pressure bottlenecks that cause frame drops on high-channel-count live views and leaves headroom for OS overhead and browser-based management consoles running concurrently.
- 1TB M.2 PCIe NVMe 2280 TLC 4X4 SSD: NVMe 4×4 (PCIe Gen 4) storage delivers sequential read bandwidth an order of magnitude faster than SATA SSDs. For a VMS workstation, this means faster database queries on event indexes, faster software launch, and lower latency on concurrent read/write during live record and playback. TLC NAND at 1TB is appropriate for OS + application + working-set storage; bulk video retention should route to a dedicated NVR or NAS.
- NVIDIA Quadro RTX A100 GPU: The Quadro RTX A100 brings professional-grade ISV certification and ECC memory to the Z2 SFF platform. This matters when running VMS software that leverages GPU decoding across high-resolution streams — NVIDIA NVDEC offloads H.265/H.264 decoding from the CPU, extending the channel count the workstation can manage in real time without frame drops.
- Intel Graphics (integrated): The 245K's on-die Intel graphics provide a no-cost display output path for headless deployments or secondary display use, preserving the discrete GPU's full bandwidth for compute and decode tasks rather than desktop rendering.
- Small Form Factor chassis (SFF): At 13.98 lb and an 8-inch width, the Z2 SFF G1i occupies the same footprint as a standard mini-tower but delivers workstation-class components. This matters in control rooms with fixed rack-adjacent furniture, reception desks with under-counter deployments, or field offices where a tower workstation is impractical.
Integration & Compatibility
The BN5L2UT#ABA runs on the Intel Z890 or equivalent 15th-gen chipset platform, making it compatible with standard enterprise imaging tools (WDS, SCCM/Intune) for rapid deployment across multi-seat rollouts. The Quadro RTX A100 supports NVIDIA's professional driver stack and ISV certifications relevant to VMS platforms that have validated GPU-decode paths. Consult your VMS vendor's hardware compatibility list to confirm GPU-accelerated decode support for your channel count and codec mix. For high-channel surveillance deployments requiring dedicated storage, pair this workstation with a professional NVR or network-attached storage rather than relying on the 1TB onboard SSD for retention. Review the full workstation category for alternative configurations matching different core-count or GPU requirements.
Frequently Asked Questions
Q: What processor is in the HP Z2 SFF G1i BN5L2UT#ABA?
A: The BN5L2UT#ABA is configured with an Intel Core Ultra 5 245K (15th Generation), a 14-core processor running at 3.60 GHz base with boost up to 5.20 GHz, plus Intel AI Boost NPU delivering up to 13 TOPS of AI inferencing capacity.
Q: What GPU does the BN5L2UT#ABA include?
A: The unit ships with an NVIDIA Quadro RTX A100 discrete GPU, which supports professional ISV driver certification and GPU-accelerated decode for video workloads alongside the integrated Intel Graphics in the Core Ultra 5 245K.
Q: Is 32GB RAM enough for running VMS software on this workstation?
A: 32GB is the practical baseline for multi-channel VMS client or server deployments. It provides headroom for the OS, VMS application, concurrent browser-based management consoles, and background agent services without triggering memory pressure on high-channel live views.
Q: Can the 1TB NVMe SSD handle video retention for a surveillance deployment?
A: The 1TB M.2 PCIe NVMe SSD is appropriate for OS, applications, and working-set storage. For sustained video retention across multiple cameras, a dedicated NVR or NAS provides the correct architecture — the onboard SSD is not designed for continuous high-bitrate write workloads over extended periods.
Q: What are the physical dimensions of the HP Z2 SFF G1i?
A: The Z2 SFF G1i measures 19.00 × 16.00 × 8.00 inches (L × H × W) and weighs 13.98 lb, making it suitable for under-desk or credenza deployment where a standard tower workstation would not fit.
Q: Does the BN5L2UT#ABA support Windows 11 Pro?
A: The Intel Core Ultra 5 245K and Z2 SFF G1i platform meet Windows 11 hardware requirements. Confirm the specific OS license included with this SmartBuy SKU at time of purchase, as SmartBuy configurations may ship with Windows 11 Pro or be available in a no-OS variant depending on the channel configuration.
The BN5L2UT#ABA stands out to me primarily because of the Intel AI Boost NPU delivering up to 13 TOPS on-die — that is dedicated silicon for AI inferencing that does not compete with your VMS decoding pipeline or your application threads. If you are running analytics-heavy VMS software on a head-end workstation and hitting CPU utilization ceilings on your current hardware, this is the spec to evaluate first.
Technical Highlights:
- Core Ultra 5 245K — 14 cores to 5.20 GHz: The hybrid core architecture schedules background OS and agent tasks on efficiency cores, leaving performance cores available for VMS decode threads and GUI responsiveness. The 5.20 GHz ceiling keeps single-stream latency low on display-side VMS clients.
- Intel AI Boost (up to 13 TOPS): The NPU handles neural-network inference workloads independently of CPU and GPU, which matters on VMS platforms that have begun offloading object detection and classification to hardware accelerators. This frees GPU bandwidth for video decode rather than competing for it.
- NVIDIA Quadro RTX A100 + NVMe 4×4 SSD: GPU-accelerated decode across H.265/H.264 streams paired with PCIe Gen 4 NVMe storage means the bottleneck shifts to your network switch and camera bitrate budget — not the workstation itself.
Deployment Considerations:
- The 1TB onboard NVMe is correctly sized for OS and application storage but is not an appropriate destination for bulk video retention — route camera streams to a dedicated NVR or NAS and use this workstation as the VMS management head-end only.
- At 8 inches wide and 13.98 lb, the Z2 SFF G1i fits in under-desk deployments, but confirm your enclosure has adequate rear clearance for the GPU and PCIe cable routing before finalizing the install location.
This configuration is the right pick for a VMS head-end or on-premise AI analytics station in a physical-security operations center where floor space is limited — not a camera storage server, but a capable management and processing node that will not become a bottleneck on analytics-enabled multi-camera deployments.
