NVIDIA RTX A6000 rental: early access
Ampere · Workstation (being confirmed)
We’re collecting early-access requests. We’re placing our capacity in Uzbekistan. We don’t publish prices yet. We’ll confirm availability in our reply.
RTX A6000
Request RTX A6000We’ll reply with terms for your workload. Nothing to pay before the contract.
Estimate usage- Memory per GPU
- 48 GB GDDR6
- Memory bandwidth
- 0.77 TB/s
- GPU interconnect
- supports an NVLink bridge for 2 GPUs
NVIDIA RTX A6000 specifications
- Architecture
- Ampere
- Announced
- 2020
- Memory
- 48 GB
- Memory type
- GDDR6
- Memory bandwidth
- 0.77 TB/s
- FP32
- 38.7 TFLOPS
- FP16/BF16, dense
- 154.8 TFLOPS
- FP8, dense
- none
- FP4, dense
- none
- FP64
- not in NVIDIA’s card specsNVIDIA’s whitepaper gives FP64 on the GA102 chip as 1/64 of FP32.
- FP64 Tensor Core
- noneNVIDIA’s whitepaper: GA102 has no FP64 Tensor Core acceleration.
- RT cores
- yes
- NVENC/NVDEC
- NVENC 1, NVDEC 2
- MIG
- none
- GPU interconnect
- supports an NVLink bridge for 2 GPUs
- Power (TDP/TGP)
- 300 W
- Form factor
- Workstation (being confirmed)
Memory is specified per GPU. Multiple GPUs have separate memory: using a larger model across them requires compatible software and model partitioning. NVLink does not make them one GPU.
The RTX A6000 form factor is still being confirmed. Specs from NVIDIA’s official materials, checked 26 September 2026.
The RTX A6000 by workload
- Inference: LLMs up to 14B in BF16 on one GPU
- Fine-tuning: QLoRA up to 32B, LoRA up to 14B on one GPU
- Training from scratch: models up to 1B on one GPU
- Rendering: scenes up to 32 GB
- FP64 computing: not suited to this GPU
An estimate from open formulas, not a benchmark or an invoice.
Servers with the NVIDIA RTX A6000
Server configuration (CPU, RAM, disk, network) is being confirmed.
Sizing assumption: up to 8 GPUs per server. We’ll confirm the actual configuration in our reply.
More than 8 GPUs? Discuss a clusterSources
- NVIDIA RTX PRO Blackwell GPU Architecture (v1.1), Table 4 (opens in a new tab)
- NVIDIA RTX A6000 Datasheet (opens in a new tab)
- NVIDIA Ampere GA102 GPU Architecture (opens in a new tab)
- NVIDIA Video Encode and Decode Support Matrix (opens in a new tab)
- NVIDIA Newsroom, November 18, 2020 (Q3 FY2021 results) (opens in a new tab)
Similar GPUs
- NVIDIA RTX 6000 Ada Generation
48 GB GDDR6 · 0.96 TB/s · PCIe 4
RTX 6000 Ada leads on 3 of 4 metrics. RTX 6000 Ada has more FP32 (91.1 vs 38.7 TFLOPS) and FP16/BF16 (364 vs 154.8 TFLOPS). Equal: memory size.
- NVIDIA L40S
48 GB GDDR6 · 0.86 TB/s · PCIe 4
L40S leads on 3 of 4 metrics. L40S has more FP32 (91.6 vs 38.7 TFLOPS) and FP16/BF16 (362.05 vs 154.8 TFLOPS). Equal: memory size.
We’ll reply with terms for your workload. Nothing to pay before the contract.