---
format: "aidr-story-markdown/v1"
id: "71d333ab8e8f0644d98a30ab30acb8e830bc8a0ec0de26224c470049efd0d1cf"
canonical_url: "https://aidr.today/71d333ab?lang=en"
title: "Exa boosts embedding throughput by 50% to index billions of documents on AWS AI chips"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-18T23:58:22.000Z"
category: "Data"
topics: ["amazon","inference","data-engineering"]
source_urls: ["https://aws.amazon.com/blogs/machine-learning/exa-boosts-embedding-throughput-by-50-to-index-billions-of-documents-on-aws-ai-chips/"]
summary: "Exa optimized embedding inference on AWS Neuron hardware using custom NKI kernels, resulting in a 47% gain in performance and cost per million documents."
---

# Exa boosts embedding throughput by 50% to index billions of documents on AWS AI chips

> [Open the canonical story](<https://aidr.today/71d333ab?lang=en>)

**Published:** 2026-11-18T23:58:22.000Z
**Category:** Data
**Topics:** amazon, inference, data\-engineering

## Summary

Exa optimized embedding inference on AWS Neuron hardware using custom NKI kernels, resulting in a 47% gain in performance and cost per million documents\.

## Sources

- [Story source](<https://aws.amazon.com/blogs/machine-learning/exa-boosts-embedding-throughput-by-50-to-index-billions-of-documents-on-aws-ai-chips/>)

