Pinterest Commits $4B to AWS Through 2031 for AI Infrastructure

Pinterest has announced its largest infrastructure investment to date: a planned $4 billion commitment to Amazon Web Services extending through 2031. The deal positions AWS as the foundation for Pinterest's next generation of AI-powered visual discovery serving more than 600 million monthly users worldwide.

The agreement expands a partnership that began in 2010, when Pinterest first adopted AWS for its cloud infrastructure. According to details first reported by Amazon, the renewed commitment will support AI model training, inference workloads, and core platform infrastructure as Pinterest accelerates its shift toward transformer-based generative models and conversational search capabilities.

Custom silicon at the center

Pinterest plans to deploy AWS Trainium chips to host and run the large language models and vision-language models that drive its personalized visual search engine. The company is also expanding its use of AWS Graviton processors, which already power approximately one-third of its compute infrastructure, to support discovery systems at scale.

The move to Amazon's custom silicon reflects a broader industry trend as companies seek alternatives to traditional GPU infrastructure for AI workloads. Pinterest's AI capabilities are built on its proprietary Taste Graph, which helps users progress from broad inspiration to specific, actionable results—whether finding recipes, shopping for products, or planning home projects.

Most recently, the platform launched Pinterest Assistant, a multi-turn conversational discovery feature powered by open-source vision-language models optimized for large-scale deployment.

"This expanded commitment with AWS gives us the compute flexibility, hardware optionality, and infrastructure efficiency to accelerate our AI vision for the next generation of visual discovery on Pinterest," said Matt Madrigal, Pinterest's chief technology officer.

Infrastructure modernization underway

Beyond AI workloads, Pinterest is using the expanded AWS relationship to modernize its underlying infrastructure. The company is migrating from traditional EC2-based environments to a Kubernetes-based architecture running on Amazon Elastic Kubernetes Service (EKS). This transition aims to improve developer velocity, operational reliability, and infrastructure efficiency across Pinterest's global platform.

The migration represents a significant architectural shift for a platform that has operated one of the largest-scale data lakes on AWS. Pinterest has evolved its AI systems from traditional retrieval methods to transformer-based models in recent years, requiring more flexible and powerful infrastructure to support these advances.

Why it matters

This deal signals how visual search platforms are betting heavily on custom AI infrastructure to differentiate their products. For Pinterest, the investment ties personalization capabilities directly to hardware choices, potentially creating competitive advantages in recommendation quality and operational costs. The commitment also validates AWS's custom silicon strategy as viable alternatives to Nvidia GPUs emerge for specific AI workloads at scale.

Dave Brown, senior vice president of AWS Compute & ML Services, noted that AWS aims to help Pinterest's teams "move faster and think bigger—benefiting users all over the world."

These details were first reported by Amazon in a company announcement about the partnership.