OpenAI skips data labelers for Goldman bankers |
![]() |
| OpenAI is sidestepping the data annotation sector by hiring ex-Wall Street bankers to train its AI models. |
| In a project known internally as Project Mercury, the company has employed more than 100 former analysts from JPMorgan, Goldman Sachs and Morgan Stanley, paying them $150 an hour to create prompts and financial models for transactions such as IPOs and corporate restructurings, Bloomberg reported. The move underscores the critical role that curating high-quality training datasets plays in improving AI model capabilities, marking a shift from relying on traditional data annotators to elite financial talent to instruct its models on how real financial workflows operate. |
| “OpenAI’s announcement is a recognition that nobody writes financial documents better than highly trained analysts at investment banks,” Raj Bakhru, co-founder of Blueflame AI, an AI platform for investment banking now part of Datasite, told The Deep View. |
| That shift has the potential to shake up the $3.77 billion data labeling industry. Startups like Scale AI and Surge AI have built their businesses on providing expert-driven annotation services for specialized AI domains, including finance, healthcare and compliance. |
| Some AI experts say OpenAI’s approach signals a broader strategy: cut out the middlemen. |
| “Project Mercury, to me, clearly signals a shift toward vertical integration in data annotation,” Chris Sorensen, CEO of PhoneBurner, an AI-automation platform for sales calls, told TDV. “Hiring a domain expert directly really helps reduce vendor risk.” |
| But not everyone sees it that way. |
| “While it’s relatively straightforward to hire domain experts, creating scalable, reliable technology to refine their work into the highest quality data possible is an important — and complex — part of the process,” Edwin Chen, founder and CEO of Surge AI, told TDV. “As models become more sophisticated, frontier labs increasingly need partners who can deliver the expertise, technology, and infrastructure to provide the quality they need to advance.” |
![]() |
| OpenAI hiring former bankers to create training data might sound like a flex, but it’s also a sign of how far it still has to go. As the pressure to monetize AI grows, the company is trying to prove its models can perform serious, industry-specific work on the job. Training them to develop financial models is a logical next step. However, paying $150 an hour for curated data highlights a deeper issue: building useful AI still requires a significant amount of manual effort. While Project Mercury may help OpenAI pitch itself to Wall Street clients, it also highlights the limitations of current models and the significant amount of hand-holding they still require to succeed. |
-
Archives
- December 2025
- November 2025
- October 2025
- September 2025
- August 2025
- July 2025
- June 2025
- May 2025
- April 2025
- March 2025
- February 2025
- January 2025
- December 2024
- November 2024
- October 2024
- September 2024
- August 2024
- July 2024
- June 2024
- May 2024
- April 2024
- March 2024
- February 2024
- January 2024
- December 2023
- November 2023
- October 2023
- September 2023
- August 2023
- July 2023
- June 2023
- May 2023
- April 2023
- March 2023
- February 2023
- January 2023
- December 2022
- November 2022
- October 2022
- September 2022
- August 2022
- July 2022
- June 2022
- May 2022
- April 2022
- March 2022
- February 2022
- January 2022
- September 2021
- August 2021
- July 2021
- June 2021
- May 2021
- April 2021
- February 2021
- January 2021
- December 2020
- November 2020
- October 2020
- September 2020
- August 2020
- July 2020
- June 2020
- May 2020
- April 2020
- March 2020
- February 2020
- January 2020
- December 2019
- November 2019
- October 2019
- September 2019
- August 2019
- July 2019
- June 2019
- May 2019
- April 2019
- March 2019
- February 2019
- January 2019
- December 2018
- November 2018
- October 2018
- September 2018
- July 2018
- June 2018
- May 2018
- April 2018
- March 2018
- February 2018
- January 2018
- December 2017
- November 2017
- October 2017
- September 2017
- August 2017
- July 2017
- June 2017
- May 2017
- April 2017
- March 2017
- February 2017
- January 2017
- December 2016
- November 2016
- October 2016
- September 2016
- August 2016
- July 2016
- June 2016
- May 2016
- April 2016
- March 2016
- February 2016
- January 2016
- December 2015
- November 2015
- October 2015
- September 2015
- August 2015
- July 2015
- June 2015
- May 2015
- March 2015
- January 2015
-
Meta

