NVIDIA A30 24GB HBM2 PCIe Gen4 Datacenter GPU
Brand: NVIDIA | Category: GPUs
SKU: 900-21001-0006-000 | Part #: 900-21001-0006-000 | MPN: 900-21001-0006-000
Contact for Pricing — Request a Quote
About the NVIDIA A30 24GB HBM2 PCIe Gen4 Datacenter GPU
The NVIDIA A30 delivers 24 GB of HBM2 memory with 933 GB/s bandwidth, purpose-built for enterprise datacenters requiring mixed-precision inference and compute workloads. Based on the NVIDIA Ampere architecture, this PCIe Gen4 GPU combines high memory capacity with exceptional memory bandwidth to accelerate demanding AI, analytics, and virtualization applications at scale.
Designed for AI infrastructure teams and datacenter operators, the A30 (part number 900-21001-0006-000) provides 3584 CUDA Cores and 224 third-generation Tensor Cores to deliver up to 10.3 TFLOPS of FP32 performance, 82 TOPS of TF32 tensor performance, 165 TOPS of BFLOAT16, and 330 TOPS of INT8 processing power. ECC memory protection ensures reliability in mission-critical environments, while Multi-Instance GPU (MIG) support enables up to 4 independent MIG instances per card, maximizing resource utilization. The dual-slot, full-height full-length form factor fits standard datacenter chassis, with passive cooling and a modest 165W thermal design power that simplifies thermal infrastructure. NVLink Bridge support allows direct GPU-to-GPU communication via NVLink 3.0 for up to 2 GPUs, enabling efficient multi-GPU applications. This NVIDIA solution operates reliably up to 83°C GPU junction temperature and requires only forced-airflow chassis ventilation, making it an ideal choice for performance-dense, cost-efficient datacenter deployments.
Key Specifications
- Memory Capacity: 24 GB HBM2
- Memory Bandwidth: 933 GB/s
- CUDA Cores: 3584
- Host Interface: PCIe Gen4 x16
- Thermal Design Power: 165W
- Manufacturer Part Number: 900-21001-0006-000
Contact Omnixon Global to request a quotation for the NVIDIA A30 24GB HBM2 PCIe Gen4 Datacenter GPU.
Technical Specifications
| Brand | NVIDIA |
| Category | GPUs |
| SKU | 900-21001-0006-000 |
| Part Number | 900-21001-0006-000 |
| Condition | New |
| Manufacturer Part Number | 900-21001-0006-000 |
| GPU Architecture | NVIDIA Ampere |
| Memory Capacity | 24 GB HBM2 |
| Memory Bandwidth | 933 GB/s |
| CUDA Cores | 3584 |
| Tensor Cores | 224 (3rd Generation) |
| FP32 Performance | 10.3 TFLOPS |
| TF32 Tensor Performance | 82 TOPS |
| BFLOAT16 Tensor Performance | 165 TOPS |
| INT8 Tensor Performance | 330 TOPS |
| Host Interface | PCIe Gen4 x16 |
| Form Factor | Dual-slot, Full-Height Full-Length (FHFL) |
| Thermal Design Power (TDP) | 165W |
| Max Operating Temperature | 83°C (GPU junction) |
| MIG Support | Yes — up to 4 MIG instances (4x MIG-6g.6gb profile equivalent) |
| NVLink Support | NVLink Bridge (2 GPUs via NVLink 3.0) |
| ECC Memory | Yes |
| Display Outputs | None |
| Cooling | Passive (requires forced airflow chassis) |
Frequently Asked Questions about NVIDIA A30 24GB HBM2 PCIe Gen4 Datacenter GPU
What server platforms accept the NVIDIA A30 24GB HBM2 PCIe Gen4 Datacenter 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.