Pinterest Commits $4 Billion to AWS Trainium and Graviton AI Chips
PINS•Pinterest will invest $4 billion in AWS AI infrastructure through 2031, tapping Trainium and Graviton chips to train and run large language and vision-language models. It follows a restructuring with layoffs of under 15% staff and is the largest infrastructure investment to advance AI-driven visual discovery for 600 million users.
1. Investment Scope
Pinterest plans to spend $4 billion on Amazon Web Services over the next five years, marking its largest-ever infrastructure commitment aimed at bolstering AI model training and inference capabilities.
2. Technology Components
Under the agreement, Pinterest will use AWS Trainium chips for high-speed machine learning training and expand use of Graviton processors, which already power about one-third of its systems, to support AI-driven features.
3. Strategic Rationale
The investment seeks to enhance personalized visual search and AI-assisted discovery for over 600 million active users by improving compute performance, hardware flexibility, and infrastructure efficiency for large language and vision-language models.
4. Context and Timeline
This commitment follows Pinterest’s global restructuring earlier this year, which included layoffs of under 15% of staff, and extends a partnership with AWS that began in 2010, with spending phased through 2031.
