Pinterest is making its biggest infrastructure investment ever with a $4 billion commitment to Amazon Web Services as it ramps up artificial intelligence capabilities for more than 600 million monthly users.
Quick Summary – TLDR:
- Pinterest has signed a $4 billion cloud services agreement with Amazon Web Services through 2031.
- The deal is the largest infrastructure commitment in Pinterest’s history.
- Pinterest will use AWS Trainium and Graviton chips to train and run AI models at scale.
- The partnership aims to improve visual search, AI assisted discovery, and advertising performance for more than 600 million users worldwide.
What Happened?
Pinterest announced on Thursday that it will spend $4 billion on Amazon Web Services cloud infrastructure through 2031, significantly expanding a partnership that has been in place since 2010. The agreement marks the largest infrastructure deal ever signed by the visual discovery platform as it seeks to accelerate AI development and support future growth.
The company plans to use Amazon’s custom AI chips, including Trainium and Graviton, to power large language models, vision language models, and other AI systems that drive personalized recommendations and discovery experiences.
📌☁️ @Pinterest just committed $4B for @awscloud services through 2031, the largest infrastructure commitment in the company’s history.
— AWS Newsroom (@AWSNewsroom) June 4, 2026
Learn how AWS is powering personalized discovery for Pinterest’s 600M+ users with Trainium, Graviton, & Amazon EKS: https://t.co/RsqTYUrGrn pic.twitter.com/g77tS06moQ
Pinterest Makes Its Biggest Infrastructure Bet Yet
The new agreement highlights how central artificial intelligence has become to Pinterest’s long term strategy. The company has increasingly invested in AI-powered tools designed to help users discover content, products, and ideas more efficiently.
According to Pinterest, the expanded AWS commitment will provide the computing resources needed to train and deploy advanced AI models while improving infrastructure efficiency and performance.
Matt Madrigal, Chief Technology Officer at Pinterest said:
The partnership builds on a relationship that began in 2010 and has helped improve the reliability and performance of Pinterest’s core services over the years.
AI Becomes a Bigger Part of Pinterest’s Platform
Pinterest has been using artificial intelligence for years through its proprietary Taste Graph, a system designed to understand user interests and deliver personalized recommendations.
The company has recently accelerated those efforts by moving from traditional retrieval systems to more advanced transformer based generative AI models. These technologies are helping Pinterest improve visual search, content recommendations, and shopping experiences.
One of the newest examples is Pinterest Assistant, a conversational discovery tool that allows users to interact with Pinterest’s search experience in a more natural way. The feature is powered by vision language models optimized for large scale deployment.
Under the AWS agreement, Pinterest plans to use Trainium to host and run large language models and vision language models that power features such as personalized visual search and AI-assisted discovery.
Amazon’s Custom Chips Take Center Stage
A major part of the deal involves Pinterest increasing its use of Amazon’s custom silicon.
The company already uses Graviton processors across roughly one third of its compute infrastructure. Pinterest now plans to expand that usage further while also leveraging Trainium to support growing AI workloads.
By adopting AWS custom chips, Pinterest expects to improve the price performance of its AI operations while reducing dependence on more expensive computing resources.
The move comes as technology companies across the industry search for cost effective ways to scale AI systems without dramatically increasing infrastructure expenses.
Infrastructure Modernization Efforts Continue
Beyond AI, Pinterest is also using the expanded AWS partnership to support a broader infrastructure modernization strategy.
The company is transitioning from traditional EC2 based environments to a Kubernetes based architecture running on Amazon Elastic Kubernetes Service (EKS).
Pinterest believes the migration will improve:
- Developer productivity
- Operational reliability
- Infrastructure efficiency
- Platform scalability
AWS executives say the partnership will help Pinterest move faster as AI adoption continues to grow across its products.
Dave Brown, Senior Vice President of AWS Compute & ML Services said:
SQ Magazine Takeaway
I think this deal shows that Pinterest is no longer treating AI as just another product feature. The company is making a massive infrastructure commitment that puts artificial intelligence at the center of its future.
While rivals like TikTok, Instagram, and Facebook continue to compete for user attention, Pinterest is betting that smarter discovery tools and personalized experiences will help it stand out. The scale of this investment suggests the company believes AI will become one of its biggest growth drivers over the next several years.
