Social media platform Pinterest has expanded its use of Amazon Web Services (AWS) with a new $4 billion infrastructure agreement.

The contract will run through 2031, building on the existing relationship the two companies have had since 2010.

Pinterest will use AWS' custom Graviton and Trainium chips under the contract. Trainium, AWS' AI chip, will be used to host and run large language models and vision-language models for Pinterest's search and AI-assisted discovery services.

Pinterest already uses Graviton for around a third of its compute infrastructure, but will increase this to run more of its systems.

In addition, Pinterest is undergoing an infrastructure modernization effort, moving from a traditional EC2 environment to a Kubernetes-based architecture, with the aim of improving reliability and infrastructure efficiency.

"Pinterest is heavily investing in AI to make discovery more personal, visual, and actionable for the hundreds of millions of people who use our platform every month," said Matt Madrigal, chief technology officer at Pinterest.

"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. This strategic partnership will help accelerate AI innovation at Pinterest, improving both our consumer experience and advertiser performance by advancing our proprietary models and our use of open-source models.”

Pinterest was "born" on AWS in 2010, and has remained a loyal customer ever since. Founded by Ben Silbermann, Paul Sciarra, and Evan Sharp, and headquartered in San Francisco, the platform is used to create digital pinboards, often focusing on recipes, home design, style, and inspiration.

Meta and Apple are known customers of Amazon's custom chips.

DCD interviewed Pinterest's former CTO in its latest magazine. Read it here