← Back to catalog
NVIDIADatacenter
NVIDIA A100 80GB PCIe
The Ampere legend, still training real models for a fraction of Hopper cost.
$12,500
80GB HBM2e
Memory
312 TFLOPS
FP16 Tensor
300W
TDP
The GPU that first proved out large-scale transformer training. form factor drops into standard servers — a proven, well-supported way to get real training throughput on a budget.
Full Specifications
Compute
- Architecture
- Ampere
- FP16 Tensor
- 312
Memory
- Memory
- 80GB HBM2e
- Memory Bandwidth
- 1.94 TB/s
Interconnect & Networking
- Interconnect
- Gen4 x16 ( bridge, 2-GPU)
Power & Cooling
- 300W
General
- Form Factor
- Dual-slot
Compatible & Required
Pairs well with