OpenAI Acquihires AI Startup Crossing Minds

While Meta poaches OpenAI talent, OpenAI has hired the team from Crossing Minds.

Crossing Minds developed an AI-powered product recommendation system for online retailers, which automatically generated shopping suggestions based on customer data. The company also offered two tools designed to ease the task of building AI applications: one for generating embeddings optimized for data like consumer shopping preferences, and another called RAGSys for equipping AI applications with retrieval-augmented generation (RAG) features.

Crossing Minds confirmed the deal, which was initially reported by TechCrunch.

The startup was launched in 2016 by a team including AI researcher Sebastian Thrun, co-founder of online education platform Udacity and Alphabet's X product research lab. The other co-founders, Emile Contal and Alexandre Robicquet, are also AI researchers. According to reports, Robicquet has joined OpenAI in a research role, where his work will focus on AI agents and post-training, a phase of AI development that enhances model output quality.

Before this development, Crossing Minds had previously raised more than $13 million in funding from investors including Shopify and Index Ventures. The current deal is a talent acquisition by OpenAI, not a new funding round for Crossing Minds.

It has not been disclosed whether OpenAI has acquired Crossing Minds’ technology or if it plans to make the technology available to its users. This marks at least the third acquisition for OpenAI since the beginning of the year. Last month, the company acquired Jony Ive’s io Products Inc. in an all-stock deal valued at $6.5 billion. Weeks prior, OpenAI reportedly spent approximately $3 billion to acquire Windsurf.

In the realm of AI-powered product recommendations, Crossing Minds competed with offerings such as Amazon Personalize, which allows developers to build real-time personalization and recommendation capabilities into their applications, and Google Cloud Recommendations AI, which leverages Google’s expertise in machine learning to provide tailored product suggestions.

For AI developer tools, particularly those focused on embeddings and retrieval-augmented generation, competitors include companies like Hugging Face, which provides open-source models and tools for various AI tasks, and Cohere, which offers large language models and enterprise-grade tools for building AI applications, including RAG capabilities.