Amazon EC2 Trn2 Instances and Trn2 UltraServers for AI/ML training and inference are now available
AWS News Blog
AWS has announced the general availability of Amazon EC2 Trn2 Instances and Trn2 UltraServers, powered by second-generation AWS Trainium2 chips for AI/ML training and inference.
- 4x faster and more powerful than first-generation Trn1 instances
- 30-40% better price performance compared to GPU-based instances
- Each Trn2 instance features 16 Trainium2 chips, 192 vCPUs, 2 TiB memory
- Trn2 UltraServers feature 64 Trainium2 chips with high-bandwidth interconnect
- Already powering large language models like Llama 3.1 and Claude 3.5 Haiku
- Available in US East (Ohio) region with capacity reservation options
The new instances are designed to support training and inference of trillion-parameter models with exceptional performance and efficiency.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
Dec 3
2024
2024
Amazon EC2 Trn2 instances are generally available
Dec 2
2025
2025
Announcing Amazon EC2 Trn3 UltraServers for faster, lower-cost generative AI training
Dec 3
2024
2024
Amazon EC2 P5en instances, optimized for generative AI and HPC, are generally available
Dec 23
2024
2024
AWS Neuron introduces support for Trainium2 and NxD Inference
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.