Lenovo ThinkSystem SR670 V3 with NVIDIA L4 24GB PCIe GPU
Brand: Lenovo | Category: GPUs
SKU: 4X97A90244 | Part #: 4X97A90244 | MPN: 4X97A90244
Contact for Pricing — Request a Quote
About the Lenovo ThinkSystem SR670 V3 with NVIDIA L4 24GB PCIe GPU
The Lenovo ThinkSystem SR670 V3 with NVIDIA L4 24GB PCIe GPU delivers exceptional inference acceleration, with peak compute of 607 TFLOPS at FP32, 1,214 TFLOPS at BF16, and 2,428 TFLOPS at FP8/INT8 precision—making it ideal for production AI workloads that demand both performance and power efficiency. This 2U rack server combines Lenovo's proven ThinkSystem architecture with NVIDIA's Ada Lovelace GPU architecture, offering 7,680 CUDA cores and 24GB of GDDR6 memory per accelerator. The dual 4th Generation Intel Xeon Scalable processor foundation, paired with up to 8TB of DDR5 system memory across 32 DIMM slots, creates a balanced platform for demanding enterprise inference, video transcoding, virtual desktop infrastructure, and professional visualization workflows.
Built for IT infrastructure teams deploying GPU-accelerated services at scale, the SR670 V3 supports up to 10 single-width PCIe GPUs via PCIe Gen 4 x16 interfaces, with enterprise-grade systems management delivered through the Lenovo XClarity Controller (XCC2). Each L4 accelerator consumes just 72W, enabling dense deployments with manageable thermal and power footprints. Hardware-accelerated encode and decode for AV1, H.264, and H.265 standards accelerate video processing pipelines. The system runs on industry-standard operating systems including Red Hat Enterprise Linux, SUSE Linux Enterprise Server, VMware vSphere, and Windows Server. For organizations seeking proven GPU acceleration in a compact, power-efficient form factor, Lenovo's ThinkSystem SR670 V3 with NVIDIA L4 24GB PCIe (part number 4X97A90244) combines enterprise reliability with modern AI infrastructure demands. Contact Omnixon Global to request a quotation for this solution.
Key Specifications
- Manufacturer: Lenovo
- Part Number: 4X97A90244
- Form Factor: 2U Rack Server
- GPU Memory: 24GB GDDR6
- CUDA Cores: 7,680
- Maximum System Memory: 8TB DDR5
- GPU Power Draw: 72W
- Maximum GPU Support: Up to 10 PCIe single-width GPUs
Technical Specifications
| Brand | Lenovo |
| Category | GPUs |
| SKU | 4X97A90244 |
| Part Number | 4X97A90244 |
| Condition | New |
| Manufacturer Part Number | 4X97A90244 |
| Product Line | ThinkSystem SR670 V3 |
| Form Factor | 2U Rack Server |
| GPU Model | NVIDIA L4 24GB PCIe |
| GPU Architecture | NVIDIA Ada Lovelace |
| GPU Memory | 24GB GDDR6 |
| GPU TDP | 72W |
| CUDA Cores | 7680 |
| Tensor Core Generation | 3rd Generation (FP8, FP16, BF16, TF32, INT8) |
| GPU Interface | PCIe Gen 4 x16 |
| Max GPU Support (Single-Width) | Up to 10 PCIe single-width GPUs |
| Processor Support | Dual 4th Generation Intel Xeon Scalable Processors |
| Memory Slots | 32 DIMM slots |
| Maximum System Memory | 8TB DDR5 |
| PCIe Generation (System) | PCIe Gen 5 |
| Systems Management | Lenovo XClarity Controller (XCC2) |
| Operating System Support | Red Hat Enterprise Linux, SUSE Linux Enterprise Server, VMware vSphere, Windows Server |
| Target Workloads | AI Inference, Video Transcoding, VDI, Professional Visualization |
| Video Encode/Decode | Hardware-accelerated AV1, H.264, H.265 encode and decode |
Frequently Asked Questions about Lenovo ThinkSystem SR670 V3 with NVIDIA L4 24GB PCIe GPU
What server platforms accept the Lenovo ThinkSystem SR670 V3 with NVIDIA L4 24GB PCIe GPU?
Reference servers include Dell PowerEdge XE9680 / XE9712, HPE Cray XD670, Lenovo ThinkSystem SR685a / SR675 V3, Supermicro AS-A21GE / SYS-821GE, Gigabyte G593 / G894, ASUS ESC. Share your target platform in the RFQ and we will confirm chassis-to-GPU compatibility and recommended NIC pairing.
How long is the lead time on AI GPUs?
Highly model-dependent. L40S / RTX-class: typically 3-6 weeks. H100/H200/B200 in SXM form factor: 12-16 weeks for whole-platform allocations. We quote genuine-channel ETAs only — no grey-market promises.
Do you supply matched networking (Quantum InfiniBand / Spectrum-X)?
Yes — Omnixon stocks the full NVIDIA networking lineup (Quantum-2 / Quantum-X InfiniBand, Spectrum-X Ethernet, ConnectX NICs, BlueField DPUs) so we can quote a complete training-cluster BOM, not just the GPUs.
Can you help with NVIDIA AI Enterprise licensing?
Yes. We hold genuine-channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.