Newly surfaced materials from the Musk v. Altman lawsuit reveal that Valve CEO Gabe Newell donated over $20 million to OpenAI in 2018 and was listed as the sole member of an informal advisory board, adding a significant and previously undisclosed dimension to the company's early governance and funding history.
The number is specific enough to be striking: $20,008,279, reported by PC Gamer from documents emerging through the ongoing litigation between Elon Musk and Sam Altman. Gabe Newell, best known as the founder of Valve and the man who controls Steam, one of the dominant platforms in PC gaming, was quietly one of OpenAI's more substantial early backers during the nonprofit era. He was not merely a donor. He was listed as the sole member of what the organization described as an informal advisory board, a governance structure so loosely defined that its existence has gone essentially unmentioned in the years of coverage that have followed OpenAI's rise to systemic importance. That combination of financial scale and structural ambiguity is the part of this story that deserves more attention than the headline number alone.
The broader cast of names appearing in the related materials underscores how much of OpenAI's early network has remained underexamined. Elon Musk and Sam Altman are the well-documented principals whose falling-out is the subject of the litigation itself. Shivon Zilis, who has been central to Musk's orbit across multiple ventures, appears in related email threads. PC Gamer's reporting also references Hideo Kojima, the celebrated game director, in the periphery of related correspondence, though the nature and significance of that connection remains less clear from the materials currently available. What emerges is a picture of early OpenAI existing not just as an AI research organization but as a node in an overlapping network of technology billionaires, gaming industry figures, and AI researchers whose relationships were being managed with the informality typical of early-stage nonprofits rather than the rigor that the organization's eventual scale would have warranted.
The context for Newell's 2018 donation matters considerably. OpenAI in 2018 was a research-focused nonprofit, publicly committed to the position that artificial general intelligence should be developed safely and its benefits distributed broadly rather than captured by any single company. The capped-profit structure that allowed commercial investment did not arrive until 2019. The ChatGPT moment that made OpenAI a household name did not arrive until late 2022. A donation of $20 million to an AI safety nonprofit in 2018 was, by the standards of that moment, an act of philanthropic enthusiasm for a speculative mission rather than a strategic investment in a company that would become one of the most commercially significant technology organizations in the world within five years.
The transformation from that nonprofit starting point to OpenAI's current position, valued at over $300 billion and generating substantial commercial revenue through enterprise contracts, API access, and consumer products, is the journey that the Musk lawsuit is fundamentally about. Musk's core allegation is that the transition from nonprofit to capped-profit structure represented a betrayal of the founding mission that early donors and supporters, including himself, understood themselves to be supporting. Whether or not that legal argument prevails, the discovery process producing it is generating a detailed archaeological record of early OpenAI's governance that was never intended for public examination. Newell's advisory board role, however informal, is part of that record, and its informality is itself significant data about how the organization was run during a period when the decisions being made were establishing precedents that would matter enormously later.
The specific phrase
Also read: AI Chatbot Subscription Scams Are Exploiting the Same App Store Channels That Legitimate Startups Depend On • The Unverified GPT-5.5 Codex Leak Is Less Important Than What It Reveals About How Founders Are Auditing AI Agents in Production • The Qwen3 27B Versus Coder-Next Debate Is Really About Whether Founders Can Trust Reddit Benchmarks to Make Infrastructure Decisions
