Pinterest CEO touts open source AI: ‘massive performance’ at low costs

On an earnings call Tuesday, Pinterest CEO Bill Ready highlighted the promise of open source AI models in helping the company lower its costs while expanding visual AI use cases.

The photo board site, which often serves as consumers’ first step on their shopping journey, leverages AI technology to power a variety of features, including its personalized recommendations, a multimedia search experience that combines text and images, ad targeting, and, most recently, product discovery with the launch of Pinterest’s AI-powered assistant.

However, investors wanted to know what Pinterest’s opportunity was in proxy commerce — AI systems that can act autonomously on behalf of users — given the rapidly changing AI landscape, and how that could impact its bottom line and growth potential.

This question has become more urgent. The company expected during the third quarter earnings announcement The weaker holiday shopping season than expected, citing the tariffs imposed by President Donald Trump and their negative impact on the home furnishings category. As a result, Pinterest’s fourth-quarter revenue is expected to range between $1.31 billion and $1.34 billion, while analysts estimate $1.34 billion on average. This news sent the stock down more than 21% on Wednesday.

Despite near-term revenue concerns, Pinterest CEO Bill Reidy pointed to ways the company can maximize its use of AI and LLMs (large language models), without significantly increasing its costs. In addition to its proprietary models that are already factored into its cost structure, Reddy said the company tests off-the-shelf leading models against open source options on a regular basis, and finds the open source models promising.

“One of the really interesting things we’re seeing is that we’re getting tremendous performance from open source models specifically for Pinterest use cases on visual AI,” Reddy told investors. “Given current market prices and token costs, in early testing, we are seeing significantly lower cost orders with similar performance using fine-tuned open source models versus leading off-the-shelf proprietary models.”

He said the company plans to move forward with several open source models for different use cases, which will come at a fraction of the cost of larger model providers.

“We feel really good about the value we’re providing to the user there, our ability to align that with monetization, our ability to control those costs and deliver that effectively,” Reddy said.

The executive also touched on other areas where Pinterest is considering how to best use AI, including agent shopping. Pinterest already offers “push-button purchasing” through its partnership with Amazon, and will wait to see if users actually want the AI ​​to “push the button for them,” Ready said.

Meanwhile, Reddy suggested that Pinterest’s biggest selling point is that it guides the user through the shopping experience — something she hopes to improve by pinterest assistant, Users accompanying the AI ​​can talk to it and ask for advice and recommendations. The company says the assistant understands the user based on their boards, collections, and saves, and how they compare to people with similar tastes.

Pinterest is also being introduced Custom panels curated by artificial intelligence, Which Reddy said combines specialized human processing with artificial intelligence.