L40 vs L40S: Specs, VRAM & Price
Short answer: pick the L40S when you're running an AI/ML workload (inference, fine-tuning), which is what the L40S is positioned and typically priced for. Pick the L40 when you're doing graphics, rendering or a mixed workload where the L40 is available cheaper.
Both the L40 and the L40S ship with 48 GB of VRAM.
You can rent both by the hour, no purchase required. L40 from $0.759/GPU/hr; L40S from $0.726/GPU/hr. The L40S is 4% cheaper at the entry rate.
L40
Good for fine-tuning mid-size models and high-throughput inference
L40S
Good for fine-tuning mid-size models and high-throughput inference
Last updated: 2026-10-10 22:00:49 UTCRefreshes hourly
Specs side by side
| Spec | L40 | L40S |
|---|---|---|
| VRAM | 48 GB | 48 GB |
| Memory bandwidth | 864 GB/s | 864 GB/s |
| Architecture | Ada Lovelace | Ada Lovelace |
| Compute capability | 8.9 | 8.9 |
| Precisions in hardware | FP32, FP16, BF16, FP8, INT4 | FP32, FP16, BF16, FP8, INT4 |
| Peak FP16/BF16 tensor TFLOPS | 181.05 TFLOPS | 362 TFLOPS |
| Peak FP8 tensor TFLOPS | 362 TFLOPS | 733 TFLOPS |
| Max board power (TDP) | 300W | 350W |
| Form factor | – | PCIe |
| Regions | 3 | 5 |
On-demand price
| Rate | L40 | L40S |
|---|---|---|
| Lowest per GPU/hr | $0.850/GPU/hr | $1.07/GPU/hr |
Which is better for AI: L40 or L40S?
The two share identical memory specs, both listing 48GB GDDR6 at 864 GB/s on NVIDIA's own product pages (Aug 2026), so this isn't a memory decision at all; the L40S is NVIDIA's data-center/AI-positioned variant of the same silicon, while the L40 is positioned for graphics/rendering workloads. Pick the L40S when you're running an AI/ML workload (inference, fine-tuning), which is what the L40S is positioned and typically priced for; pick the L40 when you're doing graphics, rendering or a mixed workload where the L40 is available cheaper. On price the L40S undercuts the L40 by 4% at the entry rate ($0.726 vs $0.759/GPU/hr). Worth weighing if the deciding factor above isn't a hard requirement for your job.
Decision criteria
| VRAM & bandwidth | Identical on paper: both are 48GB GDDR6 at 864 GB/s per NVIDIA's own L40 and L40S product pages (Aug 2026). |
| Positioning | L40S is NVIDIA's AI/data-center-optimized SKU; L40 is positioned for graphics and rendering, same silicon family but different target workload. |
| Compute precision | Both are Ada Lovelace with FP8/BF16/INT4 support (lib/tools/gpu-capabilities.ts), no precision gap. |
| Availability | 3 region(s) list the L40 vs 5 for the L40S. |
| Entry price | $0.759/GPU/hr (L40) vs $0.726/GPU/hr (L40S). |
Pick the L40 when
- You're doing graphics, rendering or visualization work
- L40 pricing is meaningfully cheaper in your region
- You have a mixed graphics+compute workload matching the L40's positioning
Pick the L40S when
- You're running inference or fine-tuning, an AI/ML-first workload
- You want the SKU NVIDIA specifically positions for data-center AI
- L40S has better availability where you're deploying
Price-performance
At list rates, 1,000 GPU-hours costs $726 on the L40S against $759 on the L40. Neither NVIDIA nor Aquanode publishes a workload-normalized $/token or $/epoch figure for this pair, so $/GPU-hour below is the only apples-to-apples number. A faster card can still cost less per finished job even at a higher hourly rate.
L40 vs L40S: common questions
Is the L40 or the L40S better for AI?
Short answer: pick the L40S when you're running an AI/ML workload (inference, fine-tuning), which is what the L40S is positioned and typically priced for. Pick the L40 when you're doing graphics, rendering or a mixed workload where the L40 is available cheaper.
Is the L40 or the L40S cheaper to rent?
On Aquanode's live marketplace the L40 starts at $0.759/GPU/hr (median $0.902/GPU/hr) and the L40S starts at $0.726/GPU/hr (median $1.20/GPU/hr). The L40S is the cheaper of the two at the entry rate, by 4%.
What is the difference between the L40 and the L40S?
Both cards report 48 GB of VRAM; both are Ada Lovelace parts; both run BF16, FP8, INT4 workloads in hardware. On price, the L40 lists from $0.759/GPU/hr and the L40S from $0.726/GPU/hr.
How widely available are the L40 and the L40S?
The L40 is listed in 3 regions and the L40S in 5 regions right now.
How much does 1,000 GPU-hours cost on the L40 vs the L40S?
At the lowest rates listed today, 1,000 GPU-hours costs $759 on the L40 and $726 on the L40S, a difference of $33 for the same runtime. Rates are per GPU per hour and update hourly.
How much VRAM does the L40 have compared to the L40S?
Both the L40 and the L40S ship with 48 GB of VRAM.
What's the actual difference between the L40 and L40S?
On paper, nothing on memory. Both report 48GB of GDDR6 at 864 GB/s on NVIDIA's own product pages. The difference is positioning: the L40S is NVIDIA's data-center/AI-workload SKU, while the L40 targets graphics and rendering. Pick based on your workload and whichever is cheaper in your region, not on a spec-sheet gap that doesn't exist.
Does the L40 or the L40S support NVLink?
Neither the L40 nor the L40S has a dedicated GPU-to-GPU interconnect.
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 L40 page and L40S 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 L40 VRAM calculator and the L40S VRAM calculator, see vendor-published throughput in the GPU benchmarks and specs table, or browse every model by generation in the GPU index.
More L40 comparisons
More L40S comparisons
Related reading: L40 pricing and specs, The NVIDIA Inception program, explained, Google Colab alternatives for dedicated GPU access, Free GPU credits for students and researchers, and RunPod volume disk vs. network volume, compared.
