Leonardo AI Drives Tech Evolution Using Generative AI on AWS

Leonardo AI began with a vision to empower game developers to build and train an artificial intelligence (AI) model using their own intellectual property to fast track their productions. In less than two years, their tools have reduced production timelines by an order of magnitude. Today, over 15 million users are able to generate content at the speed of thought through Leonardo AI, and games were just the beginning. The startup has expanded to retail, marketing, entertainment, media, and real estate applications.

However, rapid scaling requires innovative thinking. When Leonardo’s legacy architecture couldn’t handle the sudden increase in traffic, AWS was there to help. In only two weeks, the teams worked together to migrate their architecture, enabling global scale, while maintaining performance and user experience. 

Furthering their relationship with AWS, Leonardo AI participated in the first AWS Generative AI Accelerator Program, enabling them to meet with some of the world’s top generative AI startups and partners. Their participation gave them access to business and technical mentorship, as well as AWS credits, creating a significant benefit able to refine their platform and access other industry networks.

When building their solution, Leonardo used Amazon EC2 P4d Instances for both training and inference, moving to Amazon EC2 G6e, also G5 for creating their product features in real time.  Today, they use Amazon EC2 Inferentia2 instances for low cost, high performance generative AI inference while running several large language models (LLMs). Using Inferentia2, Leonardo AI reduced costs by 60% cost, while improving their LLM speed up to 35%.

Are you ready to add generative AI to your startup? Learn how you can begin using AWS Activate credits with Amazon Bedrock today.