Pinterest looks like a collection of pictures until you follow what people are doing with them. They are searching, comparing and deciding. Pinterest says that happens more than 80 billion times a month, with more than half of those searches commercial. That turns visual understanding into both an AI problem and an advertising surface.2
Pinterest says users conduct more than 80 billion searches per month, the vast majority visual and more than half commercial. The company is building shared multimodal AI infrastructure for visual understanding and introducing Visual Search Ads that let advertisers bid for prominent placement while users search and compare.1
Pinterest’s visual-search scale
The search query can be an image, not just a sentence
Pinterest says the vast majority of its searches are visual. The system has to understand objects, style and context inside images, then connect those signals with language and products. Pinterest says its NVIDIA-backed foundation gives teams shared infrastructure for that multimodal work.1
Commercial intent makes the AI layer monetizable
A search for a bag, chair or outfit can sit close to a purchase decision even without a conventional product query. Pinterest’s Visual Search Ads place advertiser responses in search results and Pin closeups, where people are already comparing what they see.2
Pinterest says advertisers can bid for a top response and send people directly to their sites. The product turns visual understanding into ad inventory: recognize the intent, find a relevant commercial response and sell placement at that moment.2
The infrastructure and ad product reinforce each other
Better multimodal understanding can make discovery more relevant. More commercial searches create more opportunities to monetize that discovery. The business loop is stronger when the same AI foundation supports user experience and advertising rather than living as a separate research project.
This resembles Shopify’s effort to make product data legible to AI from the other side. Shopify starts with merchant catalogs. Pinterest starts with consumer visual intent. Both connect machine understanding with a transaction path.
Eighty billion searches is not eighty billion purchases
The scale figure describes search activity, and the commercial-share figure describes intent as Pinterest defines it. Neither is a conversion count or revenue figure. A commercial search can end without a click or purchase, and the new ad format still has to prove its economics for advertisers.
The business chain to watch
- Visual input gives Pinterest a distinctive intent signal.
- Multimodal models turn images and language into searchable meaning.
- Commercial searches create moments where advertisers may be relevant.
- Visual Search Ads sell placement inside those moments.
Pinterest’s AI story is easy to mistake for a model-infrastructure announcement. The more interesting business story is what sits underneath it: tens of billions of searches, a large commercial-intent pool and a new way to sell access to it.
Sources and methodology
Sources checked September 25, 2026. Dates and periods for individual figures are stated beside them.
- Pinterest: multimodal AI foundation with NVIDIA ↗Accessed 2026-09-25
- Pinterest: Visual Search Ads announcement ↗Accessed 2026-09-25
Scope and assumptions
Pinterest-reported search volume and commercial-intent share are company metrics, not independent measurements of purchases or advertiser returns.
Visual Search Ads are newly announced; the article does not infer conversion lift, revenue contribution or return on ad spend.
Continue reading
AI Search Orders on Shopify Are Up Nearly 13×. Is Chat Becoming a Storefront? →