Lenovo 30J5002VUS ThinkStation P3 Ultra Small Form Factor Workstation
Overview
The Lenovo ThinkStation P3 Ultra SFF G2 (30J5002VUS) is a compact professional workstation built around Intel's Core Ultra 5 235 processor and an NVIDIA RTX A400 discrete GPU — a pairing that gives security integrators and surveillance architects genuine GPU-accelerated decode and analytics processing inside a chassis that fits on a shelf or under a desk. At 22 lb and 19 × 11 × 8.80 inches, it ships enough compute headroom for multi-stream video analytics without the footprint of a tower workstation. If you're sourcing a VMS server for a mid-size deployment — 32 to 64 channels of analytics-heavy streams — this SFF package is worth a close look before defaulting to rack-mounted hardware.
Lenovo's ThinkStation workstation line has a long track record in surveillance control rooms, and the P3 Ultra SFF G2 targets the segment where space and power budget matter as much as raw throughput. Explore the full range of professional workstations to compare tower and rack-mount alternatives if your deployment demands greater expansion capacity.
Key Features
- Intel Core Ultra 5 235 — 14 Cores at up to 4.40GHz: The hybrid 14-core architecture (performance + efficiency cores) handles concurrent workloads that a traditional quad-core would stall on — running a VMS service, a GPU analytics engine, and remote desktop sessions simultaneously without throttling. The 24MB cache keeps frequently accessed frame buffers close to the compute units, which matters when decoding high-bitrate H.265 streams in real time.
- NVIDIA RTX A400 4GB GPU: This is the spec that separates the P3 Ultra from a generic business PC on a surveillance workstation shortlist. The RTX A400 provides hardware-accelerated video decode across H.264 and H.265 — offloading stream decode from the CPU entirely. For VMS platforms like Milestone XProtect or Genetec Security Center that support GPU decode, this translates directly into more concurrent decoded streams per server and lower CPU utilization headroom for analytics tasks.
- 32GB DDR5-6400 (2× SO-DIMM, expandable to 128GB): DDR5 at 6400MHz doubles the theoretical bandwidth of DDR4-3200, which matters when the RTX A400 and the CPU are both pulling frame data from system memory simultaneously. Starting at 32GB with two slots occupied means expansion requires replacing both modules — plan your memory ceiling at procurement if you anticipate growth to 64GB or 128GB, since there's no free slot to add into.
- 512GB PCIe Gen5 M.2 2280 TLC Opal SSD: PCIe Gen5 delivers sequential read speeds roughly double those of Gen4, which keeps OS, VMS application, and database I/O from becoming a bottleneck on startup or during incident export. The Opal self-encrypting drive (SED) layer means the drive can be hardware-encrypted without software overhead — important for deployments that require data-at-rest encryption for compliance without sacrificing read/write performance. The 512GB drive is the OS and application volume; plan separate NAS or NVR storage for video retention.
- Intel BE200vPro Wi-Fi 7 (802.11be): Wi-Fi 7 support via the Intel BE200vPro card provides up to 5.8Gbps theoretical throughput on 6GHz bands — overkill for most VMS servers, but practically useful in temporary deployment environments or mobile command posts where wired Ethernet isn't available. For permanent installations, wire it: Ethernet remains the correct network interface for a fixed surveillance server.
- Bluetooth 5.4: Peripheral connectivity for keyboards, mice, and presentation devices without a USB dongle consuming a port. Not a primary spec for a VMS server, but removes one friction point in a control room setup.
- 230W Internal Power Supply: The 230W PSU is sized for the RTX A400 and Core Ultra 5 under sustained load. This is a single-PSU configuration — not redundant. For 24/7 critical surveillance infrastructure, factor in a UPS on the circuit. The 230W draw also means standard 15A office circuits handle it comfortably without dedicated power drops.
- Windows 11 Pro 64-bit (Pre-installed): Ships with Windows 11 Pro — the OS baseline required by most enterprise VMS platforms. No additional OS licensing cost or reinstall needed before provisioning your VMS software. The 'Pro' license also enables BitLocker, domain join, and Remote Desktop — standard requirements for enterprise security deployments.
- Small Form Factor Chassis: The SFF enclosure fits into security control room furniture, under-desk deployments, and rack shelves with a shelf adapter. At 22 lb, a single technician can carry and install it without assistance. The compact footprint also reduces cooling load compared to a full tower, which matters in closet or furniture-enclosed installations.
Integration & Compatibility
The 30J5002VUS ships with PCIe, USB, and Ethernet interfaces — the three physical connectivity standards that matter for VMS integration. The RTX A400 GPU is a supported hardware decode accelerator in Milestone XProtect, Genetec Security Center, Avigilon Control Center, and other major VMS platforms that expose GPU decode APIs. Confirm your VMS vendor's current GPU compatibility list before procurement, as accelerated decode support is version-dependent. The Intel BE200vPro is a Wi-Fi 7 and Bluetooth 5.4 combo card — it occupies an M.2 slot and does not consume a PCIe expansion slot.
For surveillance deployments, pair this workstation with a dedicated network video recorder or NAS device for video storage — the 512GB internal SSD is not a practical video retention volume for any multi-camera deployment. A structured approach to PoE switch selection alongside the workstation will simplify the network architecture for camera-to-server connectivity. The SO-DIMM memory configuration and SFF chassis limit internal expansion; if your deployment needs dual GPUs or more than two storage drives internally, step up to the ThinkStation P5 or P7 tower family.
Frequently Asked Questions
Q: What is the maximum RAM supported by the 30J5002VUS?
A: The ThinkStation P3 Ultra SFF G2 supports up to 128GB of DDR5-6400 memory across its two SO-DIMM slots (2× 64GB modules). It ships with 32GB (2× 16GB), so upgrading requires replacing both installed modules — there is no free slot available for a simple add-on.
Q: Does the RTX A400 support hardware-accelerated video decode for VMS platforms?
A: Yes. The NVIDIA RTX A400 provides hardware H.264 and H.265 decode acceleration. VMS platforms including Milestone XProtect and Genetec Security Center expose GPU decode APIs that the RTX A400 supports. Check your specific VMS version's GPU compatibility documentation to confirm the RTX A400 is on the validated hardware list for your release.
Q: Is the SSD in the 30J5002VUS self-encrypting?
A: Yes. The 512GB M.2 2280 PCIe Gen5 SSD is a TLC Opal drive, meaning it supports hardware-level self-encryption compliant with the TCG Opal 2.0 standard. This allows BitLocker (included with Windows 11 Pro) to use the drive's hardware encryption engine rather than software encryption, avoiding any performance penalty.
Q: What is the warranty on the 30J5002VUS?
A: The 30J5002VUS includes a 3-year onsite manufacturer warranty, as indicated in the full Lenovo product name for this configuration.
Q: Is the 30J5002VUS suitable as a standalone VMS server for a 32-camera deployment?
A: It depends on your stream resolution, frame rate, and analytics load. The Core Ultra 5 235 with RTX A400 GPU decode offload provides meaningful headroom for mid-size deployments. However, the 512GB internal SSD is not adequate for video retention at that scale — you will need external NAS or NVR storage. For GPU-accelerated decode of 1080p or lower streams, this platform is appropriately sized; for 4K-heavy or deep-learning analytics deployments, evaluate the P5 or P7 tower platforms.
Q: Does the 30J5002VUS have a redundant power supply?
A: No. The SFF chassis includes a single 230W internal power supply. For 24/7 critical surveillance infrastructure, connect it to an appropriately sized UPS to protect against power interruptions.
The 30J5002VUS is the workstation I'd reach for when a client needs GPU-accelerated VMS decode in a space-constrained control room environment. The RTX A400 4GB paired with the Core Ultra 5 235's 14-core architecture gives you the combination that actually matters for surveillance workloads: dedicated NVENC/NVDEC silicon offloading stream decode so the CPU stays available for analytics, database writes, and remote access sessions running concurrently.
Technical Highlights:
- PCIe Gen5 NVMe SSD: Sequential reads at Gen5 speeds (roughly 10-14GB/s depending on drive model) mean VMS database queries, incident export, and OS operations don't compete for I/O bandwidth — a common bottleneck on Gen3/Gen4 workstations running high-channel-count VMS deployments.
- DDR5-6400 at 32GB (expandable to 128GB): The 6400MHz memory bus gives the RTX A400 and Core Ultra 5 the bandwidth headroom to handle simultaneous GPU decode and CPU analytics without memory becoming the constraint. Ceiling at 128GB means this platform can grow well beyond its base configuration — relevant if your VMS licensing or analytics load expands.
- 230W Single PSU in SFF Chassis: Lenovo has sized the power supply appropriately for the RTX A400 under sustained GPU decode load. The 230W draw fits a standard 15A circuit without requiring a dedicated power drop — important in retrofit installations where the electrician isn't part of the project scope.
Deployment Considerations:
- The two SO-DIMM slots are both populated at 2× 16GB from the factory. Any memory upgrade means replacing both modules, not adding to them — spec your memory requirement upfront rather than planning an incremental upgrade path.
- The 512GB internal SSD is strictly OS and application storage. No surveillance deployment at meaningful channel count should use this drive for video retention — connect to external NAS, a dedicated NVR, or a SAN from day one.
For a mid-size physical security operations center — 20 to 50 cameras, mixed 1080p and 4K, running Milestone XProtect or Genetec with GPU decode enabled — the ThinkStation P3 Ultra SFF G2 fits the compute and spatial requirements cleanly. It is not the right answer for AI-heavy 4K-everywhere deployments or sites requiring hot-swap redundant power; for those, step to a rack-mounted platform.
