Analysis
OpenAI posted a job listing for an investment-banking expert -- paying up to $205,000 a year plus equity -- specifically tasked with improving ChatGPT's ability to handle M&A and fundraising-related work, according to The Register. The listing lands alongside a separate, larger effort: OpenAI has hired more than 100 former bankers from Goldman Sachs, JPMorgan and Morgan Stanley at roughly $150 an hour for a project internally code-named "Mercury," training AI models specifically on research, valuation, financial modeling and diligence work.
The combination is a more direct signal than most AI-and-jobs stories: OpenAI isn't just building a general-purpose model that happens to be useful for finance, it's deliberately recruiting the exact professionals whose junior-level work -- comps, models, diligence memos -- it's training the model to replicate. That's a meaningfully different posture than earlier waves of "AI could disrupt finance" commentary, which tended to be speculative rather than backed by a named internal project and a specific hiring pipeline from the banks themselves.
Goldman Sachs has been unusually candid about this risk publicly, with executives previously acknowledging AI's potential to compress junior-analyst headcount over time. OpenAI hiring directly from Goldman, JPMorgan and Morgan Stanley to build the tools that could eventually replace some of that same junior work adds a pointed, almost adversarial dimension -- the banks' own former employees are effectively training their replacement, for pay, at the same firm racing to commercialize it.
“This lands in the same week as Layoffs.fyi data showing AI has now been cited in nearly 88,000 layoffs in 2026, 22% of the year's total.”
This lands in the same week as Layoffs.fyi data showing AI has now been cited in nearly 88,000 layoffs in 2026, 22% of the year's total. Finance and banking haven't been the largest single contributor to that tally so far, but OpenAI's targeted investment-banking push suggests the sector's exposure is about to become more direct and more measurable, not less.
For VCs and fintech founders, the practical read is that AI-native alternatives to traditional junior-banker work -- automated diligence, modeling, comp analysis -- are moving from a startup thesis (Rogo, Hebbia and similar tools have been pitching this for two years) to something a frontier lab is now building in-house, which changes the competitive landscape for any startup selling AI-for-finance tools directly to banks. For LPs and finance-sector operators, the practical question isn't whether this compresses junior headcount eventually -- most agree it will -- but how fast, and which firms adapt their staffing models proactively versus reactively.
What to watch next: whether OpenAI packages this work into a dedicated finance-vertical product rather than folding it into general ChatGPT capability, and whether other banks respond by building competing in-house tools rather than depending on OpenAI's.
