skip to main content


The AWS re:Invent 2018 event held in Las Vegas in early December saw a wide range of product announcements in the artificial intelligence (AI) and machine learning (ML) space, ranging from new services to enhancements. Most significant was the availability, slated for 2019, of a custom Amazon chip for ML inference mode, named AWS Inferentia.


  • Amazon's acquisition of chipmaker Annapurna pays returns with the 2017 launch of "Nitro" chips.

Features and Benefits

  • Learn about the latest features for Amazon SageMaker, such as automated labeling of AI training data.
  • Discover Amazon's new inference mode chip called Inferentia.

Key questions answered

  • Which type of AI user is Amazon SageMaker targeted at?
  • What are the cloud economics of using GPUs on AWS?

Table of contents

Ovum view

  • Summary
  • Amazon's acquisition of chipmaker Annapurna pays returns
  • Amazon introduces auto-labeling for ML training data
  • Amazon SageMaker targets ML developers and data scientists
  • The economics of cloud GPUs on AWS


  • Author