OpenAI’s secretive Project Mercury recruits over 100 former Wall Street analysts to train its AI on financial modeling and deal workflows. Hinting at a future where AI could replace much of the grunt work done by junior bankers.
Automating the Analyst Grind
OpenAI is quietly building an AI system that could automate the tedious, time-consuming work of junior investment bankers. The company’s internal project, known as Project Mercury, reportedly involves more than 100 former banking analysts from major firms like JPMorgan Chase, Goldman Sachs, and Morgan Stanley.
These ex-bankers are being paid about $150 an hour to teach the AI how to handle typical analyst duties — from complex Excel models to pitch-deck formatting and market analysis.
Why It Matters
If successful, Project Mercury could automate 60–70% of routine financial tasks, potentially reshaping how banks train and employ new analysts.
Critics warn that relying on AI for early-career work may erode human judgment and weaken the skill development traditionally built through hands-on experience.
For OpenAI, the project offers access to high-value financial workflows — giving its models an edge in enterprise and fintech markets where accuracy, speed, and structure matter most.
What to Watch
- Whether investment banks begin adopting AI trained under Project Mercury.
- The impact on entry-level roles and career paths in finance.
- Regulatory reactions if AI systems start handling sensitive financial data.
- How competitors like Anthropic or Google DeepMind respond in the financial-AI space.
