Most of OpenAI’s talent pipeline originates from Stanford and University of California.
According to Harmonic data obtained by Rubén Domínguez, the top 10 universities that feed the pipeline of workers who develop LLMs such as GPT-5 come from institutions like Stanford, UC, MIT, Carnegie, Harvard, among others; a list entirely consisting of U.S. universities.
This raises a crucial point: The networks and communities that form between students working on technologies like AI help new ideas move faster from research to real-world use. It is no coincidence that part of the founding and early team of companies such as OpenAI or Anthropic have ties to these institutions; in Europe, you see links like Hugging Face and Mistral with École Polytechnique and OpenAI’s research collaboration with the University of Oxford.
In sum, the institutional origin of talent matters: it influences research, the use of that research, and the technical standards that, in turn, shape the trajectory of technology. Understanding these networks is key for countries and universities that seek to develop their own capabilities and reduce gaps.