On November 5, 2025, during Pinterest’s earnings call, CEO Bill Ready emphasized the significant advantages of integrating open-source artificial intelligence (AI) models into the company’s operations. He highlighted that these models not only enhance performance but also substantially reduce costs, particularly in the realm of visual AI applications.
Pinterest, a platform renowned for inspiring users in their shopping journeys, leverages AI to power features such as personalized recommendations, multimodal search experiences combining text and images, and targeted advertising. The recent introduction of the AI-driven Pinterest Assistant further underscores the company’s commitment to enhancing product discovery through advanced technology.
Investors have expressed interest in understanding Pinterest’s position within the rapidly evolving AI landscape, especially concerning agentic commerce—AI systems capable of autonomously acting on behalf of users. This inquiry has gained urgency following Pinterest’s third-quarter earnings announcement, which projected a weaker-than-expected holiday shopping season. The company attributed this forecast to the impact of President Donald Trump’s tariffs on the home furnishings sector, leading to an anticipated fourth-quarter revenue between $1.31 billion and $1.34 billion, slightly below analysts’ average estimate of $1.34 billion. This news resulted in a stock decline of over 21% on November 6, 2025.
Despite these short-term financial challenges, Ready outlined strategies to maximize AI and large language models (LLMs) usage without significantly increasing expenses. He noted that, in addition to proprietary models already integrated into Pinterest’s cost structure, the company regularly evaluates leading off-the-shelf models against open-source alternatives. The findings have been promising, with open-source models delivering exceptional performance tailored to Pinterest’s visual AI needs.
One of the really, really interesting things that we’re seeing is that we are just getting tremendous performance from open source models specifically for Pinterest use cases on visual AI, Ready informed investors. He elaborated that, given current market rates and per-token costs, early testing has demonstrated a significant reduction in expenses with comparable performance when using fine-tuned open-source models versus leading proprietary alternatives.
Ready further stated that Pinterest plans to adopt numerous open-source models for various applications, achieving comparable performance at a fraction of the cost associated with larger model providers.
In addition to discussing AI integration, Ready addressed Pinterest’s approach to agentic shopping. He acknowledged that while the company offers push-button type buying through its partnership with Amazon, the concept of AI autonomously making purchases on behalf of users is still in its early stages. Ready emphasized that Pinterest’s primary focus remains on guiding users through their shopping experiences, enhancing this journey with tools like the Pinterest Assistant. This AI companion understands users’ preferences based on their boards, collages, saves, and similarities to others with comparable tastes.
Furthermore, Pinterest is experimenting with AI-powered personalized boards, combining expert human curation with AI-driven suggestions. These boards aim to provide users with tailored content, including trending styles, weekly outfit inspirations, and shoppable items, thereby enriching the overall user experience.
In summary, Pinterest’s strategic adoption of open-source AI models reflects a commitment to innovation and cost efficiency. By leveraging these technologies, the company aims to enhance its platform’s capabilities, offering users a more personalized and engaging experience while maintaining financial prudence.
