Leaked Documents Show OpenAI Has a Clear Definition of 'AGI' (gizmodo.com)

Leaked internal documents reported by The Information reveal that OpenAI and Microsoft quietly agreed in 2023 on a commercial definition of “AGI”: an AI system would be deemed AGI if it generates at least $100 billion in profits. The deal stems from Microsoft’s roughly $13+ billion investment in OpenAI and includes a profit-sharing arrangement that would entitle Microsoft (and other investors) to returns until they recoup $100 billion, after which OpenAI would cut off Microsoft’s access to new technology. OpenAI is currently far from that threshold (projected ~ $4B revenue in 2024), so under the current terms the company would likely remain contractually tied to Microsoft for a long time. This framing is significant because it ties the technical milestone of “outperforming humans at most tasks” to an economic metric, shifting AGI from a capability-based safety/governance issue to a commercial one. The documents illuminate strategic friction: OpenAI’s push to restructure as for‑profit, Microsoft’s interest in equity or terminating exclusivity on cloud hosting, Google’s antitrust concerns to the FTC, and legal pressure from Elon Musk. For the AI/ML community, the leak highlights how governance, investor incentives, and corporate contracts can shape research priorities, access to models, competition, and the distribution of power—arguably as consequential as the underlying technical advances themselves.