NVIDIA H20 141GB HBM3e
Hopper SXM 2025 4nm
FP32
44.0
TFLOPS
VRAM
141 GB
HBM3e
TDP
400 W
110.0 TFLOPS/kW
Bandwidth
4.0 TB/s
memory
Performance Metrics
Peak theoretical throughput by precision type
| Precision | Bits | Peak TFLOPS | Efficiency |
|---|---|---|---|
| INT8 | 8 | 296.0 | 0.740 TFLOPS/W |
| FP8 | 8 | 296.0 | 0.740 TFLOPS/W |
| FP16 | 16 | 148.0 | 0.370 TFLOPS/W |
| BF16 | 16 | 148.0 | 0.370 TFLOPS/W |
| TF32 | 32 | 74.0 | 0.185 TFLOPS/W |
| FP32 | 32 | 44.0 | 0.110 TFLOPS/W |
| FP64 | 64 | 1.0 | 0.003 TFLOPS/W |
FP8 Efficiency
0.740 TFLOPS/W
296.0 TFLOPS / 400W
FP16 Efficiency
0.370 TFLOPS/W
148.0 TFLOPS / 400W
FP32 Efficiency
0.110 TFLOPS/W
44.0 TFLOPS / 400W
Power Specifications
TDP
400 W
Max Power
460 W
Power Connector
PCIe 16-pin
Cooling
Liquid
Memory Specifications
Capacity
141 GB
Type
HBM3e
Bandwidth
4000 GB/s
Interface
--
Hardware & Design
Form Factor
SXM
Architecture
Hopper
Process Node
4nm
Launch Year
2025
Variant
141GB HBM3e
Market Segment
Data Center
Full Specifications
| Memory | |
|---|---|
| VRAM | 141 GB |
| Memory Type | HBM3e |
| Bandwidth | 4.0 TB/s |
| Interconnect & I/O | |
| GPU-to-GPU | NVLink |
| Interconnect Bandwidth | 900 GB/s |
| Power & Thermal | |
| TDP | 400 W |
| General | |
| Form Factor | SXM |
| Architecture | Hopper |
| Launch Year | 2025 |
Documentation & Resources
Common Use Cases
General Compute AI/ML Workloads Data Processing
The NVIDIA H20 is optimized for high-performance computing tasks with Hopper architecture delivering 44 TFLOPS of compute power.
Where to Rent
Compare cloud providers offering on-demand GPU instances for AI training, inference, and HPC workloads.
Browse GPU Cloud ProvidersSimilar GPUs
Stay Updated on GPU Releases
Get notified when new GPUs are added or specifications are updated.
Loading verification...
No spam, unsubscribe anytime.