NVIDIA A100 80GB PCIe

Compare NVIDIA A100 80GB PCIe against another machine in the Arena

Reference: Ampere (microarchitecture) on Wikipedia

Class: GPU
Memory: 80 GB HBM2e
Runs (Q4_K_M): —
Bandwidth: 1935 GB/s
TDP: 300 W
Released: 2021-06-01
Price (US): $5500 used · $10000 new as of 2026-08
Price (Canada): CA$20319 new sourced, as of 2026-08

Sourcing and disambiguation notes

Distinct from the nvidia-a100-80gb entry, which documents the SXM4 module (2039 GB/s, 400W, requires an HGX baseboard). This PCIe card uses the same 80GB HBM2e but a narrower memory clock, giving 1935 GB/s at 300W, per NVIDIA's PCIe product brief. Pricing (Jarvis Labs, updated 2026-04-19): new $8,000-$12,000, used $4,000-$7,000, midpoints recorded here. A100 prices have generally been falling as enterprises move to H100/H200, so treat as a rough range, not a precise quote. The Canadian figure is a retail list price, near the top of the US broker range rather than its midpoint -- consistent, not anomalous.

Specification sources

https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/a100/pdf/PB-10577-001_v02.pdf — vendor, 2022-03-01
https://gpudojo.com/a100-80gb — press, 2026-08-22

Measurements

Decode is token generation — the speed you feel while an answer streams. Prefill is prompt processing — the wait before it starts. Why bandwidth predicts decode speed.

What verified, single-source and estimated mean, and the same rows with every filter and sort in the benchmarks explorer.

Speed over time

What it can run

Fit is arithmetic, not a measurement — how it is computed. Against 80 GB; models that fit are listed largest first.