Trainium2
Trainium2 is an AWS custom AI chip (ASIC) based on the NeuronCore v3 architecture, first released in 2024.
Built from BenchGecko data as of April 14, 2026 · updates when the data changes
Trainium2 is an AWS custom AI chip (ASIC) based on the NeuronCore v3 architecture, first released in 2024.
Basic
The AWS Trainium2 is a custom AI chip (ASIC) made by AWS. Peak throughput is 667 TFLOPS at FP16/BF16 and 1,334 TFLOPS at FP8. It carries 96 GB of HBM3 with 2.9 TB/s of memory bandwidth. Rated power (TDP) is 500 W. It is built on N5 at TSMC.
Deep
In the BenchGecko dataset it sits in the frontier tier: current-generation flagship silicon shipping at scale to hyperscalers. Disclosed buyers include Anthropic and Amazon. Specs come from manufacturer datasheets; the spec card on this page shows the dataset values and their as-of date.
Expert
That is 1.33 FP16 TFLOPS per watt. Memory bandwidth per unit of compute is 4.35 GB/s per FP16 TFLOP: a higher ratio helps memory-bound inference, while peak TFLOPS matter more for compute-bound training. Fabrication: Taiwan. Real throughput depends on the model, precision, batch size and software stack, so compare tokens per second per dollar for your own workload.
Depending on why you're here
- ·Trainium2: 667 TFLOPS FP16/BF16 · 1,334 TFLOPS FP8
- ·Process: N5 at TSMC
- ·Tier: frontier
- ·96 GB of HBM3 per chip
- ·Cloud availability in the dataset: AWS
- ·Compare it with other chips on /hardware/trainium2
- ·Disclosed buyers: Anthropic and Amazon
- ·AWS · fabbed at TSMC
- ·Supply tightness is tracked on /hardware
- ·Trainium2 is a chip built to run AI, made by AWS
- ·It has 96 GB of fast memory so large models fit
- ·Companies buy or rent these to train and serve AI models
Frequently Asked Questions
Read the primary sources
- AWS Trainium2aws.amazon.com