NVIDIA · Ampere · 2020
NVIDIA A100 80GB. Rental by the hour, on demand.
From $1.73/hr per GPU on demand, or prepay one GPU at $1.69/GPU-hr for 3 months, or an 8-GPU machine at $1.60/GPU-hr for 6 months.
A100 80GB on demand
- Availability
- Available now · 6 regions
- By region and size
- $1.73 to $4.70/hr
- Prepaid
- 3 mo $1.69/hr6 mo $1.60/hr12 mo $1.54/hr
- Billing
- By the second on most GPUs, a few by the minute
- 8+ GPUs or a month
- Launch any size yourself, or prepay a machine for 1 to 12 months at a discount. Custom quotes on request.
The machine
80 GB of HBM2e. At up to 2.0 TB/s.
- Memory
- 80 GB
- HBM2e
- Bandwidth
- up to 2.0 TB/s
- Memory to compute
- FP16 tensor
- 312
- TFLOPS
- Per node
- 8 GPUs
- up to 400 W each
Price across clouds
NVIDIA A100 80GB, cloud by cloud. Sorted by price.
On-demand rates per GPU-hour from each cloud's public price page on 8 Nov 2026: RunPod's Secure Cloud, and Lambda's lowest per-GPU rate across its cluster sizes. Runcrate's row is today's live starting rate.
Workloads
What teams run on the A100 80GB. Each one opens the launch form, sized.
Launch it
Ask your agent for an A100 80GB. Or read the call it makes.
Claude Code or Codex runs it for you: it starts the instance, copies your files over and runs the job. The SDKs make the same request.
Connect your agent first
Add Runcrate's MCP server to yourself: run curl -fsSL "https://runcrate.ai/mcp.md" and follow it.
Start an A100 80GB on Runcrate called my-gpu, copy train.py to it and run it.
create_instance { name: "my-gpu", gpu: "A100_80G" }instance_status { id: "my-gpu" }file_upload { id: "my-gpu", content: "<train.py>", path: "/root/workspace/train.py" }ssh_execute { id: "my-gpu", command: "nohup python /root/workspace/train.py > train.log 2>&1 &" }Spec sheet
How the A100 80GB compares. Against its closest alternatives.
Benchmarks
The A100 80GB, measured. Real runs, dated, with what they cost.
- A100 vs H100 vs H200
One script on six GPUs (run 2026-10-25)
- GPU startup time
Launch to first CUDA matmul (run 2026-10-26)
Alternatives
Alternatives to the A100 80GB. Same account, different price and performance.
| GPU | Memory · FP16 tensor | Stands out | Price gap | From |
|---|---|---|---|---|
| NVIDIA A100 80GBThis GPU | 80 GB HBM2e · 312 TFLOPS | $1.73/hr | ||
| NVIDIA A100 40GB | 40 GB HBM2 · 312 TFLOPS | -40 GB memory | +15% | $1.99/hr |
| NVIDIA L40S | 48 GB GDDR6 · 362 TFLOPS | -32 GB memory | -2% | $1.69/hr |
| AMD MI250X | 128 GB HBM2e · 383 TFLOPS | +48 GB memory | Not in stock right now | |
| NVIDIA RTX 5090 | 32 GB GDDR7 | -48 GB memory | -46% | $0.93/hr |
Full specs
NVIDIA A100 80GB specifications.
Memory
- Memory
- 80 GB
- Type
- HBM2e
- Bandwidth
- up to 2.0 TB/s
Compute
- FP32
- 19.5 TFLOPS
- FP16 tensor
- 312 TFLOPS
- INT8
- 624 TOPS
- Tensor cores
- 432
- CUDA cores
- 6,912
Power and form
- TDP
- up to 400 W
- Architecture
- Ampere
- Released
- 2020
Node
- GPUs per node
- 8
- Maker
- NVIDIA
Questions
Renting the A100 80GB. What buyers ask first.
Other GPUs
Browse every GPU by architecture.
NVIDIA Blackwell
NVIDIA Hopper
NVIDIA Ampere
NVIDIA Ada Lovelace
NVIDIA Volta