OpenAI published today a direct reply to Elon Musk's court filing, accusing him of cutting and taking private entries from Greg Brockman's journal out of context to build a different narrative. What really happened in 2017 and why does it matter today? I'll explain it clearly and without jargon.
What OpenAI says
According to OpenAI, Musk did participate in conversations about changing the organization's structure in 2017. The discussion wasn't binary: it wasn't 'non-profit or nothing'; the idea was to keep a non-profit and create a linked for-profit entity tied to the mission. That hybrid structure is basically what exists today: a PBC (public benefit corporation) and a nonprofit foundation that controls part of the PBC's equity, which OpenAI values at roughly 130 billion dollars.
OpenAI claims Musk pushed for absolute control, talked about raising $80 billion for a city on Mars, and pressed for greater stake and rights. Negotiations ended because the cofounders refused to give him total control. After several conversations, Musk left OpenAI in February 2018 and, OpenAI says, told them they could continue without him — while also saying he didn't think they'd succeed without his financial backing.
OpenAI also accuses Musk of using organizational resources to benefit his commercial projects (for example, asking for help with Tesla's Autopilot) and of launching a public and legal harassment campaign when xAI emerged.
Key points of OpenAI's account
-
In 2017 there were talks about converting part of OpenAI into a
for-profittied to the mission; Musk participated and appeared to agree with a hybrid solution. -
The cofounders preferred a B-corp with safeguards for the mission, but they also considered keeping the
non-profitif it could raise enough capital. -
Negotiations broke down because the founders refused to give Musk absolute control; he even proposed merging OpenAI with Tesla.
-
Musk stepped away in 2018; OpenAI kept looking for funding (they even considered an ICO, for example) and ended up creating the current structure:
PBC+ controlling foundation. -
OpenAI calls Musk's current lawsuit the fourth version of similar claims and sees it as part of a strategy to slow OpenAI and favor xAI.
What's really at stake?
This isn't just a personal fight between founders. Three big issues overlap in this conflict:
-
Governance: who controls a technology that can change society? You want a single founder with concentrated power, or structures with limits and a public-purpose mission?
-
Funding: can an ambitious AGI-style project survive on donations, or does it need
for-profitvehicles to attract massive capital? -
Public and legal narrative: journal excerpts and emails can be read many ways. OpenAI says Musk trims context to favor his version.
Why should you care about this?
Because this shapes how powerful technologies are built and governed. Do we trust key decisions to one person with near-total power? Or to structures that try to balance returns and public benefit? The answer affects who develops AI, under what rules, and with what incentives.
Also, public fights and lawsuits eat time and attention. That can slow collaboration between companies, scare off partners, or even push away talent who prefer less litigious environments. Ever wondered why startups sometimes avoid high-profile conflicts? This is exactly why.
What's next
This is part of an ongoing legal process and a wider war of narratives among very powerful actors. Courts, full evidence, and public hearings will help complete the picture, but we already see collateral damage: reputation hits, distraction, and increased regulatory scrutiny.
If anything is clear, it's that the story the public sees is often a cut-out piece. How do you decide who to believe when both sides use private records to justify their version? The practical lesson: check context, look for original sources, and remember that AI governance is as much political as it is technical.