Via cnet.com
The AI company has hired over 100 former investment bankers for an internal project aimed at automating financial modeling workflows
Sam Altman almost became a banker. During his sophomore year at Stanford, the OpenAI CEO says he was “peer pressured and tricked” into accepting an internship at Goldman Sachs. He ultimately turned it down, choosing a path that led him through Y Combinator and eventually to running the most influential AI company on the planet.
Now, in a twist that’s almost too neat, the company Altman built is pulling talent directly from the trading floors and deal rooms he avoided. OpenAI has recruited over 100 former investment bankers from firms including Goldman Sachs, JPMorgan Chase, and Morgan Stanley for an internal initiative called Project Mercury, which aims to bake sophisticated financial modeling capabilities into its AI systems.
Project Mercury and the $150-an-hour bankers #
The initiative, first reported by Bloomberg in October 2025, is essentially OpenAI’s attempt to teach its AI how to do the work that keeps junior bankers chained to their desks until 2 a.m. Financial modeling, the bread and butter of entry-level Wall Street life, is the core target.
The former bankers aren’t building AI from scratch. They’re serving as subject matter experts, feeding their institutional knowledge into OpenAI’s systems so the models can learn to replicate tasks that traditionally require years of spreadsheet-stained training. For this privilege, they’re compensated at roughly $150 per hour.
Altman’s banking road not taken #
Altman’s reflection on his near-miss Goldman internship adds a personal layer to what is otherwise a corporate strategy story. He’s described the idea of working in traditional banking as sounding “unbelievably terrible” compared to the startup world he ultimately chose.
Altman has led OpenAI as CEO since 2019, following his stint running Y Combinator, the startup accelerator that helped launch companies like Airbnb and Stripe.
What this means for Wall Street #
The implications of Project Mercury are straightforward and uncomfortable for anyone building a career in entry-level finance. If AI can reliably perform financial modeling, the demand for analysts who spend their first two years doing exactly that work drops considerably.
Banks themselves have been investing heavily in automation tools. JPMorgan has built its own internal AI systems. Goldman Sachs has publicly discussed using generative AI to assist with code generation and research. The difference is that OpenAI is building a general-purpose tool that could be deployed across the entire industry, not just within one firm’s walls.
For the banks, this creates an odd strategic tension. Their former employees are now helping build a product that could reduce headcount at their old employers. The knowledge transfer is flowing in one direction, out of the banks and into OpenAI’s training pipeline. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our