During an earnings call this week, Pinterest CEO Bill Ready emphasized the potential of utilizing open source AI models to help the company control costs while expanding the scope of visual AI applications.
This article highlights the strategic use of open source AI models by Pinterest to enhance visual AI capabilities and manage costs effectively.
The popular image-centric platform relies on AI technology to power various features such as personalized recommendations, multimodal search experiences, ad targeting, and the recent introduction of the AI-powered Pinterest Assistant for product discovery.
Investors expressed interest in Pinterest’s potential in agentic commerce, especially in light of the evolving AI landscape and its implications on financial performance and growth prospects.
Following a weaker holiday sales forecast due to external factors like tariffs, Pinterest anticipates fourth-quarter revenue slightly below analysts’ estimates, leading to a significant drop in stock value.
Despite revenue concerns, CEO Bill Ready outlined the company’s strategy to leverage open-source AI models and large language models (LLMs) to optimize performance without substantial cost escalation.
Ready emphasized the efficiency and cost-effectiveness of open-source models tailored for Pinterest’s visual AI use cases, signaling a shift towards leveraging such models at a fraction of the cost compared to proprietary solutions.
Techcrunch event
San Francisco
|
October 13-15, 2026
Ready also discussed Pinterest’s exploration of AI applications in agentic shopping, highlighting the platform’s existing capabilities for streamlined purchasing through partnerships like Amazon, while considering user preferences for autonomous AI-driven actions.
Pinterest continues to enhance user experience with the AI-powered Pinterest Assistant, offering personalized advice and recommendations based on individual preferences and interests.
Moreover, Pinterest introduces personalized boards curated through a blend of human curation and AI technologies, aiming to provide a unique and tailored browsing experience for users.
