AI hardware
AWS Inferentia2
FP16 throughput
1.9×10¹⁴ FLOP/s
Memory
29.8 GB
AWS Inferentia2 is an AI accelerator designed by Amazon AWS, released in April 2023.
It delivers 1.9×10¹⁴ FLOP/s of tensor FP16/BF16 throughput, 1.9×10¹⁴ FLOP/s at FP8. It carries 29.8 GB of memory at 764 GB/s of bandwidth.
Full record
- Manufacturer
- Amazon AWS
- Release date
- Apr 13, 2023
- Tensor FP16/BF16
- 1.9×10¹⁴ FLOP/s
- FP8
- 1.9×10¹⁴ FLOP/s
- FP32
- 4.8×10¹³ FLOP/s
- INT8
- 3.8×10¹⁴ OP/s
- Memory
- 29.8 GB
- Memory bandwidth
- 764 GB
More from Amazon AWS
SourceEpoch AI, 'Machine Learning Hardware'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/machine-learning-hardware. Licensed under CC BY 4.0.
← All ai hardware