Ampere
NVIDIA A40
wins
1
The NVIDIA A40 offers 48GB GDDR6 at a lower price point than the A100.
Blackwell
NVIDIA B200 180GB
wins
4
The B200 is NVIDIA's Blackwell flagship — 4.5 PFLOPS FP16 and 8 TB/s memory bandwidth make it the most powerful GPU available for AI training at scale.
Performance comparison
Visual bars — winner highlighted
NVIDIA A40
Metric
NVIDIA B200 180GB
FP16 TFLOPS
Bandwidth
VRAM
TDP (lower=better)
TFLOPS/watt
GB/s per watt
Full specification table
| Spec | NVIDIA A40 | NVIDIA B200 180GB |
|---|---|---|
| Architecture | Ampere | Blackwell |
| VRAM | 48 GB GDDR6 | 180 GB HBM3e |
| VRAM type | GDDR6 | HBM3e |
| Memory bandwidth | 696 GB/s | 8,000 GB/s |
| FP16 TFLOPS | 150 TFLOPS | 4,500 TFLOPS |
| TDP | 300 W | 1,000 W |
| TFLOPS/watt | 0.50 T/W | 4.50 T/W |
| GB/s per watt | 2.32 GB/s/W | 8.00 GB/s/W |
| NVLink | No | Yes |
| Release year | 2020 | 2024 |
Live cloud pricing
On-demand pricing across providers — updated in real time
NVIDIA A40
No live pricing data available.
NVIDIA B200 180GB
Power efficiency analysis
TFLOPS/watt and GB/s/watt — critical for data center TCO
NVIDIA A40
NVIDIA B200 180GB
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.
Performance-per-dollar
TFLOPS/$ and GB/s/$ at current best cloud prices
NVIDIA A40
NVIDIA B200 180GB
TFLOPS/$ = FP16 TFLOPS ÷ best on-demand $/hr. Higher is better — more compute per dollar. GB/s/$ = memory bandwidth ÷ best on-demand $/hr. For memory-bound inference workloads (most LLM serving), GB/s/$ is often a better predictor of cost-efficiency than TFLOPS/$. Prices are live from the Compute Comparison API and update in real time.
Workload Cost Comparison
Total job cost = best on-demand $/hr × GPU count × hours
| Workload | Scale | NVIDIA A40 | NVIDIA B200 180GB |
|---|---|---|---|
| 7B LLM fine-tune | 8× GPU, 24 h | — | $672 |
| 70B LLM fine-tune | 64× GPU, 72 h | — | $16k |
| Inference server | 1× GPU, 730 h | — | $2.6k |
| Image gen batch | 4× GPU, 8 h | — | $112 |
Based on current best on-demand prices. Spot pricing typically 40–70% lower. Actual cost depends on provider, region, and availability.
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 B200 180GB for:
- Next-gen LLM training
- Trillion-parameter models
- AI factories
Weaknesses / Limitations
- Extremely limited availability — primarily reserved for hyperscalers
- 1,000W TDP requires specialized power and liquid cooling
- Highest cost per GPU of any currently available card
Popular GPU comparisons