OpenAI’s secret ‘Project Mercury’ could replace entry-level bankers and consultants

OpenAI

OpenAI is quietly working on a major new initiative aimed at automating some of the most time-consuming financial and consulting tasks — a move that could reshape Wall Street’s junior workforce. The company has reportedly hired more than 100 former bankers and consultants from top firms including Goldman Sachs, JPMorgan Chase, and Morgan Stanley for a confidential project internally codenamed “Mercury.”

According to The Wall Street Journal, Project Mercury is designed to train AI models to handle advanced financial modelling, valuation analysis, and other complex tasks typically performed by junior investment banking analysts. These include building IPO and leveraged buyout models, conducting restructuring analyses, and preparing client pitch decks — all of which demand precision, speed, and consistency.

Sources say OpenAI is paying these ex-bankers around $150 per hour to help fine-tune its financial AI systems. Consultants involved in the project are responsible for writing prompts, reviewing the accuracy of outputs, and providing expert feedback to align results with real-world corporate standards. This iterative process is helping OpenAI build an AI that can think and respond like an experienced analyst.

A threat to entry-level banking roles?
The move is part of OpenAI’s broader push into enterprise solutions, positioning the company in direct competition with global consulting firms such as McKinsey, BCG, and Deloitte. By developing AI tools capable of performing financial due diligence and data-driven decision support, OpenAI could soon automate tasks that typically keep thousands of junior analysts busy each year.

The initiative follows growing adoption of OpenAI’s enterprise models by financial institutions. For instance, Morgan Stanley has already integrated OpenAI’s technology into its wealth management division, allowing advisers to generate client insights and investment recommendations more efficiently. Early pilots under Project Mercury are reportedly being tested with select banks and investment firms to evaluate real-world accuracy and compliance.

The long-term impact
Project Mercury’s timing is notable. OpenAI recently secured a $4 billion credit facility from JPMorgan Chase and other major lenders, signalling deepening ties between the AI company and Wall Street. Ironically, JPMorgan itself is racing to become the world’s first “AI-powered megabank,” building in-house systems that rival OpenAI’s own offerings.

Industry experts say the project could transform the economics of investment banking by cutting costs and reducing reliance on entry-level human analysts. Rather than eliminating jobs outright, it’s expected to shift how financial professionals work — from manual spreadsheet building to interpreting and validating AI-generated models.

If successful, OpenAI’s Project Mercury may not just streamline workflows; it could redefine what it means to work in high finance in the AI era.