A30 vs L40S
A30 starts at $0.35/GPU-hour on-demand; L40S starts at $0.97/GPU-hour on-demand. Compare the exact variants, memory and published specifications below. Peak compute figures do not establish application speed.
A30
L40S
Spec sheet
| A30 | L40S | |
|---|---|---|
| Spec variant | A30 | L40S |
| Architecture | Ampere | Ada Lovelace |
| VRAM (spec variant) | 24 GB | 48 GB |
| FP16 TFLOPS (dense) | 165 | 362 |
| Memory bandwidth | 933 GB/s | 864 GB/s |
| TDP | 165 W | 350 W |
| Interconnect | NVLink 200GB/s | PCIe Gen4 |
| Launched | 2021 | 2023 |
Rental price per manufacturer-rated FP16 compute: A30 $2.12 per PFLOP-hour vs L40S $2.68. Formula: cheapest $/hr ÷ dense FP16 PFLOPS (1 PFLOP = 1000 TFLOPS). This uses A30 and L40S, whose datasheets may differ from the variants above. It is not measured job throughput. Specs: A30 manufacturer, L40S manufacturer.
What they cost to rent (2026-10-02)
| A30 | L40S | |
|---|---|---|
| Cheapest on-demand $/hr | $0.35 | $0.97 |
| Cheapest spot/community $/hr | – | $0.79 |
| Cheapest source | Massed Compute | Massed Compute |
| Providers offering it | 1 | 16 |
Lowest recorded on-demand rate per configuration; highlighted = cheaper. Spot/marketplace tiers shown on the individual pages.
FAQ
A30 is cheaper: the A30 rents from $0.35 per GPU-hour and the L40S from $0.97 on-demand, updated 2026-10-02.
The largest recorded A30 variant has 24 GB and the largest L40S variant has 48 GB. Other variants may have less memory. The L40S has more.
Compare more rental options: the screener · price movers · methodology.