Skip to main content

NVIDIA L40S vs NVIDIA L4: Spec Comparison & Buyer's Guide

In AI infrastructure selection, NVIDIA L40S and NVIDIA L4 are two accelerators frequently compared. This article contrasts them item by item — architecture, compute, memory, power, and release cadence — to help you quickly judge which fits training or inference workloads.

Spec Comparison Table​

VendorNVIDIA L40SNVIDIA L4
VendorNVIDIANVIDIA
ArchitectureAda Lovelace AD102Ada Lovelace AD104
ProcessTSMC 4NTSMC 4N
Release Date2023 82023 3 GTC
FP8 Compute733 TFLOPS485 TFLOPS
FP16 Compute——
FP32 Compute91.6 TFLOPS30.3 TFLOPS
INT8 Compute——
Memory Type——
Memory Capacity48 GB GDDR6 ECC24 GB GDDR6
Memory Bandwidth864 GB/s300 GB/s
TDP Power350 W72 W

Key Differences​

  • FP8 compute: NVIDIA L40S leads with ~733 TFLOPS versus NVIDIA L4's 485 TFLOPS, a clear edge in large-scale Transformer training/inference.
  • Power: NVIDIA L4 has a TDP of 72 W, lower than NVIDIA L40S's 350 W, friendlier to datacenter PUE and cooling.
  • Memory capacity: NVIDIA L40S packs 48 GB GDDR6 ECC, more than NVIDIA L4's 24 GB GDDR6, more comfortable for single-card hosting of very large models.

Selection Advice​

  • When chasing extreme single-card compute and a mature toolchain, prioritize NVIDIA L40S; if budget, power wall, or local support are hard constraints, NVIDIA L4 often fits better. Use this site's AI Compute Card Comparison Tool to validate multiple chips side-by-side before deciding.

FAQ​

What are the main differences between NVIDIA L40S and NVIDIA L4?​

The core difference is architecture and compute density: NVIDIA L40S uses Ada Lovelace AD102, FP8 ~733 TFLOPS, memory 48 GB GDDR6 ECC; NVIDIA L4 uses Ada Lovelace AD104, FP8 ~485 TFLOPS, memory 24 GB GDDR6. See the comparison table above.

What is the TDP (power) of NVIDIA L40S?​

NVIDIA L40S has a TDP of 350 W; actual whole-system power also includes board, fans, and PUE.

Which is better for large-model training / inference?​

Training values memory capacity, bandwidth, and multi-card interconnect; inference values single-card throughput and power efficiency. Combine the "Key Differences" and "Selection Advice" above with your batch size, model size, and SLA.

How much do NVIDIA L40S and NVIDIA L4 differ in memory capacity?​

NVIDIA L40S is 48 GB GDDR6 ECC, NVIDIA L4 is 24 GB GDDR6; the gap directly affects loadable model size and context length.