# AWS reports AssemblyAI inference compute savings of 70 percent using Spot instances

_Published Friday, November 13, 2026 at 2:45 PM EST · AI · Latest · Tier 2 — Notable_

![AWS reports AssemblyAI inference compute savings of 70 percent using Spot instances — Primary](https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/06/ML-20884-featured-image.png)

AssemblyAI achieved 70 percent savings on inference compute compared with On-Demand pricing by using GPU Spot instances in production, AWS reported. Its speech-to-text platform maintained service commitments while compute demand grew nearly 4x year-over-year.

The system combines an always-on On-Demand baseline with spare GPU capacity spread across three instance families. When AWS reclaims a Spot instance with a 2-minute warning, unfinished transcription jobs return to a queue for another worker to process. Independent regional deployments keep audio within required geographic boundaries. Interruptions can increase processing time, and the approach suits short-lived jobs that can be retried independently.

## Sources

- [Artificial Intelligence](https://aws.amazon.com/blogs/machine-learning/cost-optimizing-assemblyai-speech-to-text-inference-on-aws-with-amazon-ec2-spot/)

---
Canonical: https://techandbusiness.org/newswire/gJUkGYPXC8CUduE_vvnY39
Published: 2026-11-13T19:45:20.156Z
Story chronology: 2026-11-13T19:35:35.000Z
Retrieved: 2026-11-13T23:13:45.021Z
Publisher: Tech & Business (techandbusiness.org)