A40 vs B200: Specs, VRAM & Price
Short answer: choose the B200 when 180 GB of VRAM is the gating requirement. Choose the A40 when 45 GB is enough for the workload and the lower hourly rate points that way.
The A40 has 45 GB of VRAM; the B200 has 180 GB.
You can rent both by the hour, no purchase required. A40 from $0.801/GPU/hr; B200 from $9.35/GPU/hr. The A40 is 91% cheaper at the entry rate.
A40
Good for training and inference on models that don't need FP8 kernels
B200
Good for training and serving frontier-scale LLMs
Last updated: 2026-11-03 UTCRefreshes every 6 hours
Specs side by side
| Spec | A40 | B200 |
|---|---|---|
| VRAM | 45 GB | 180 GB |
| Memory bandwidth | – | 8 TB/s |
| Architecture | Ampere | Blackwell |
| Compute capability | 8.6 | 10.0 |
| Precisions in hardware | FP32, FP16, BF16, INT4 | FP32, FP16, BF16, FP8, INT4 |
| Peak FP16/BF16 tensor TFLOPS | – | 2,250 TFLOPS |
| Peak FP8 tensor TFLOPS | – | 4,500 TFLOPS |
| NVLink / multi-GPU interconnect | – | NVLink 5, 1.8 TB/s per GPU |
| Max board power (TDP) | – | Up to 1,000W |
| Form factor | – | SXM |
| Regions | 1 | 1 |
On-demand price
| Rate | A40 | B200 |
|---|---|---|
| Lowest per GPU/hr | $0.801/GPU/hr (cheapest overall) | $9.35/GPU/hr |
Which is better for AI: A40 or B200?
Choose the B200 when 180 GB of VRAM is the gating requirement. Choose the A40 when 45 GB is enough for the workload and the lower hourly rate points that way.
A40 vs B200: common questions
Is the A40 or the B200 better for AI?
Short answer: choose the B200 when 180 GB of VRAM is the gating requirement. Choose the A40 when 45 GB is enough for the workload and the lower hourly rate points that way.
Is the A40 or the B200 cheaper to rent?
On Aquanode's live marketplace the A40 starts at $0.801/GPU/hr (median $0.801/GPU/hr) and the B200 starts at $9.35/GPU/hr (median $9.35/GPU/hr). The A40 is the cheaper of the two at the entry rate, by 91%.
What is the difference between the A40 and the B200?
The A40 has 45 GB of VRAM against the B200's 180 GB; the A40 is an Ampere part and the B200 is Blackwell (compute capability 8.6 vs 10.0); only the B200 supports FP8 compute. On price, the A40 lists from $0.801/GPU/hr and the B200 from $9.35/GPU/hr.
How widely available are the A40 and the B200?
The A40 is listed in 1 region and the B200 in 1 region right now.
How much does 1,000 GPU-hours cost on the A40 vs the B200?
At the lowest rates listed today, 1,000 GPU-hours costs $801 on the A40 and $9,350 on the B200, a difference of $8,549 for the same runtime. Rates are per GPU per hour and update hourly.
How much VRAM does the A40 have compared to the B200?
The A40 has 45 GB of VRAM; the B200 has 180 GB.
How this comparison is calculated
Every price is normalized to a per-GPU hourly rate using the same pipeline as every other pricing surface on Aquanode, and only the cheapest qualifying offer per provider is shown. A dash means that provider does not currently list that GPU.
Architecture, compute capability and supported precisions come from each vendor's published datasheet for that generation, not from the marketplace feed. This page regenerates at most once per hour.
Memory bandwidth, tensor TFLOPS, NVLink/interconnect and max board power are the same verified-from-datasheet figures shown on each model's own A40 page and B200 page, which link the primary vendor source for that model. A dash here means Aquanode hasn't verified that figure from a primary source yet, never a guess.
Go further
Check what fits in each card's memory with the A40 VRAM calculator and the B200 VRAM calculator, see vendor-published throughput in the GPU benchmarks and specs table, or browse every model by generation in the GPU index.
More A40 comparisons
More B200 comparisons
Related reading: H100 pricing and specs, H100 vs H200, A100 vs H100, H100 and H200: SXM vs NVL vs PCIe, and The best GPUs for AI, ranked.
