SERVCHIP
Interactive view
ARCHITECTURE
Ampere
SERIES
A100 SXM
Memory
80GB HBM2e
Bandwidth
2 TB/s
TDP
400W
In StockA100 SXM·Best for: High-Performance AI Training (SXM)
NVIDIA A100 SXM Tensor Core GPU
Ampere
|Data Center GPUsOverview
The NVIDIA A100 SXM Tensor Core GPU is the high-power SXM variant of the Ampere-architecture A100, featuring 80GB of HBM2e memory with 2 TB/s bandwidth and a 400W TDP. Designed for maximum AI training throughput, it supports NVLink 3.0 interconnect at 600 GB/s and Multi-Instance GPU (MIG) partitioning for multi-tenant deployments. The A100 SXM remains one of the most widely deployed data center accelerators in cloud and enterprise AI infrastructure worldwide.
Key Features
- 80GB HBM2e memory with 2 TB/s bandwidth
- Ampere architecture with 3rd-gen Tensor Cores
- Multi-Instance GPU (MIG) with up to 7 instances
- NVLink 3.0 interconnect at 600 GB/s
- SXM4 form factor for maximum performance
Ideal Use Cases
ai trainingai inferencehpcdata analytics
Full Specifications
Complete technical specifications for the A100 SXM.
Memory
- Memory
- 80GB HBM2e
- Memory Bandwidth
- 2 TB/s
Compute
- Tensor Cores
- 432 (3rd-gen)
- CUDA Cores
- 6912
- FP8 TFLOPS
- 624
- FP16 TFLOPS
- 312
- TF32 TFLOPS
- 156
- FP64 TFLOPS
- 19.5
Interconnect
- Interconnect
- NVLink 3.0, 600GB/s
Physical
- TDP
- 400W
- Form Factor
- SXM4
- Cooling
- Liquid Cooled
Manufacturing
- Launch Date
- Q2 2020
- Manufacturing Process
- 7nm
