Ampere
NVIDIA A40
wins
2
The NVIDIA A40 offers 48GB GDDR6 at a lower price point than the A100.
Ada Lovelace
NVIDIA L4
wins
3
The NVIDIA L4 is the most power-efficient Ada Lovelace GPU at just 72W TDP.
Performance comparison
Visual bars — winner highlighted
NVIDIA A40
Metric
NVIDIA L4
FP16 TFLOPS
Bandwidth
VRAM
TDP (lower=better)
TFLOPS/watt
GB/s per watt
Full specification table
| Spec | NVIDIA A40 | NVIDIA L4 |
|---|---|---|
| Architecture | Ampere | Ada Lovelace |
| VRAM | 48 GB GDDR6 | 24 GB GDDR6 |
| VRAM type | GDDR6 | GDDR6 |
| Memory bandwidth | 696 GB/s | 300 GB/s |
| FP16 TFLOPS | 150 TFLOPS | 242 TFLOPS |
| TDP | 300 W | 72 W |
| TFLOPS/watt | 0.50 T/W | 3.36 T/W |
| GB/s per watt | 2.32 GB/s/W | 4.17 GB/s/W |
| NVLink | No | No |
| Release year | 2020 | 2023 |
Live cloud pricing
On-demand pricing across providers — updated in real time
Power efficiency analysis
TFLOPS/watt and GB/s/watt — critical for data center TCO
NVIDIA A40
NVIDIA L4
TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.
When to choose each GPU
Choose NVIDIA A40 for:
- Large model inference
- Professional rendering
- Multi-tenant AI
Weaknesses / Limitations
- Older Ampere architecture — lower FP16 TFLOPS than L40S at same VRAM
- GDDR6 bandwidth (696 GB/s) lower than HBM-based GPUs
- No NVLink on A40
Choose NVIDIA L4 for:
- Efficient inference
- Video transcoding
- Edge AI
Weaknesses / Limitations
- 24 GB VRAM limits model size — cannot run 34B+ models
- Low memory bandwidth (300 GB/s) — bottleneck for throughput-sensitive inference
- Not suitable for training or fine-tuning beyond small models
Popular GPU comparisons