---
format: "aidr-story-markdown/v1"
id: "111c7ae55acd3d24bd5a00b49f79a28442b5d34c99a1ec12cc833f31f2348d49"
canonical_url: "https://aidr.today/111c7ae5?lang=en"
title: "How Deepgram optimized voice AI at scale with AWS Trainium"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-19T16:23:22.000Z"
category: "Infra"
topics: ["inference"]
source_urls: ["https://aws.amazon.com/blogs/machine-learning/how-deepgram-optimized-voice-ai-at-scale-with-aws-trainium/"]
summary: "Deepgram collaborated with the AWS Forward Deployed Engineering team to port its Nova-3 speech-to-text model to AWS Trainium2 for offline batch inference, covering the AWS Neuron stack, kernel-level tuning, a data-parallel Rust runtime, and throughput and cost results on batch transcription."
---

# How Deepgram optimized voice AI at scale with AWS Trainium

> [Open the canonical story](<https://aidr.today/111c7ae5?lang=en>)

**Published:** 2026-10-19T16:23:22.000Z
**Category:** Infra
**Topics:** inference

## Summary

Deepgram collaborated with the AWS Forward Deployed Engineering team to port its Nova\-3 speech\-to\-text model to AWS Trainium2 for offline batch inference, covering the AWS Neuron stack, kernel\-level tuning, a data\-parallel Rust runtime, and throughput and cost results on batch transcription\.

## Sources

- [Story source](<https://aws.amazon.com/blogs/machine-learning/how-deepgram-optimized-voice-ai-at-scale-with-aws-trainium/>)