Flopper.io Flopper.io
vs Rebellions logo

NVIDIA L40S 48GB vs Rebellions ATOM-Max

The L40S delivers 2.8x the FP16 throughput of the ATOM-Max (362 vs 128 TFLOPS dense). The ATOM-Max counters with 16 GB more memory (64 GB vs 48 GB).

L40S: Ada Lovelace, 2023 ATOM-Max: ATOM, 2024
FP16 dense lead
2.8x
L40S: 362 vs 128 TFLOPS
Memory
48 vs 64 GB
16 GB more for the ATOM-Max
Bandwidth
864 GB/s vs 1.0 TB/s
ATOM-Max moves data faster
TDP
350 W vs 350 W
Same power budget
See rental pricing from $0.53/hr Download comparison sheet

Share this page

NVIDIA L40S 48GB vs Rebellions ATOM-Max

A post with this link shows the card on the Image tab as its preview.

Post on X Share on LinkedIn
 Printable version →

NVIDIA L40S 48GB

Leads on FP16 Full specs →

What is the NVIDIA L40S?

The L40S is a GPU from NVIDIA on the Ada Lovelace architecture, an accelerator for AI training, inference and high-performance computing. It is a PCIe add-in card, so it fits standard servers and workstations. L40S full specifications →

Can I rent the L40S in the cloud?

Yes. Cloud providers tracked by Flopper rent the L40S by the hour, and the pricing page compares their current rates and the price history side by side. L40S rental prices →

How does the L40S compare with other accelerators?

This page sets the L40S against the ATOM-Max. The L40S is the older generation of the two. This is a cross-vendor matchup, NVIDIA against Rebellions, and both ship as a PCIe card, so they fit the same class of system. Flopper also compares it head to head with the TPU v5e 16GB, across every published precision, memory and power. L40S vs TPU v5e 16GB →

Where do the L40S figures come from?

Every L40S figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Rebellions logo

Rebellions ATOM-Max

Full specs →

What is the Rebellions ATOM-Max?

The ATOM-Max is an NPU from Rebellions on the ATOM architecture, an accelerator built for AI inference. It is a PCIe add-in card, so it fits standard servers and workstations. ATOM-Max full specifications →

Can I rent the ATOM-Max in the cloud?

Flopper is not tracking any live rental listings for the ATOM-Max yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the ATOM-Max compare with other accelerators?

This page sets the ATOM-Max against the L40S. The ATOM-Max is the newer generation of the two. This is a cross-vendor matchup, Rebellions against NVIDIA, and both ship as a PCIe card, so they fit the same class of system. Flopper also compares it head to head with the Rebel100, across every published precision, memory and power. ATOM-Max vs Rebel100 →

Where do the ATOM-Max figures come from?

Every ATOM-Max figure on this page traces to Rebellions's own datasheet or specification table. How Flopper verifies specifications →

Specifications

L40S
Rebellions logo
ATOM-Max
Architecture
Ada LovelaceATOM
Launch Year
20232024
Form Factor
PCIePCIe
Memory
48 GB64 GB
Memory Bandwidth
864 GB/s1.0 TB/s
TDP
350 W350 W
Process Node
4nm—

Performance (TFLOPS)

L40S
Rebellions logo
ATOM-Max
1.4 TFLOPS No verified data available
92 TFLOPS No verified data available
183 TFLOPS
366 TFLOPS sparse
 No verified data available
362 TFLOPS
733 TFLOPS sparse
 No verified data available
362 TFLOPS
733 TFLOPS sparse
128 TFLOPS
733 TFLOPS
1,466 TFLOPS sparse
 No verified data available
FP6
No verified data available No verified data available
No verified data available No verified data available
733 TOPS
1,466 TOPS sparse
512 TOPS

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The ATOM-Max is the newer part: ATOM (2024) against the L40S's Ada Lovelace (2023).

The ATOM-Max carries 16 GB more memory (64 GB against 48 GB), so larger models fit on one GPU before you shard. Bandwidth is 864 GB/s against 1.0 TB/s, which sets the ceiling on decode speed.

Power is 350 W against 350 W, and at FP32 the L40S does more work per watt. Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP64, FP32, TF32, BF16, FP16, FP8, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

Is the L40S faster than the ATOM-Max?

At FP16 precision the L40S reaches 362 TFLOPS dense against 128 TFLOPS for the ATOM-Max. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the L40S or the ATOM-Max?

Choose the L40S for training and high-throughput inference: it has 2.8x the FP16 throughput of the ATOM-Max. Choose the ATOM-Max when the model has to fit on one GPU: it carries 16 GB more memory.

Which has more memory, the L40S or the ATOM-Max?

The ATOM-Max carries 64 GB of memory versus 48 GB for the L40S. Memory bandwidth is 864 GB/s for the L40S and 1.0 TB/s for the ATOM-Max.

How much power do the L40S and the ATOM-Max draw?

The L40S is rated at 350 W TDP and the ATOM-Max at 350 W. On FP32 throughput per watt, the L40S is the more efficient part.

How much do the L40S and the ATOM-Max cost to rent?

Live cloud listings tracked by Flopper price the L40S from $0.53 per GPU-hour across 25 providers, and no live listings yet for the ATOM-Max. The rental section on this page links to every provider, rate and the price history for each GPU.

Where to Rent

NVIDIA L40S 48GB

Rent from

$0.53 / GPU-hr

Lowest of 44 live listings across 25 providers · median $1.50

See all L40S rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

Rebellions logo

Rebellions ATOM-Max

No live rental listings tracked for the ATOM-Max yet.

Browse GPU cloud providers

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

Open in Full Compare Tool → L40S full specs ATOM-Max full specs Estimate cluster cost in the calculator →

© 2026 Flopper.io - Compare the GPUs Powering AI