OpenAI fires back at Musk with receipts from 2017

When a boardroom breakup heads to court, the deadliest weapon isn't a high-priced attorney. It's the receipts sitting in old email threads.

OpenAI has fired back at Elon Musk's latest lawsuit by publishing a detailed timeline accompanied by internal emails from 2015 to 2018. Their core message is remarkably blunt: Musk is suing them for transitioning into a for-profit entity, but he was actually the first one to push for it.

Equity demands, Tesla mergers, and early doubts

According to OpenAI's disclosed timeline, Musk questioned the non-profit model before OpenAI was even formally unveiled. In a November 2015 email to Sam Altman, Musk argued that a pure non-profit would muddy incentives, suggesting it might be better to have a standard C corp with a parallel non-profit.

The real friction began in early 2017. As researchers made headway training AI on the video game Dota, they realized compute demands were skyrocketing. Ilya Sutskever noted at the time that while compute needs were ballooning exponentially, AGI could likely be built with under $10 billion in hardware.

By summer 2017, both sides agreed that a for-profit vehicle was inevitable to raise the billions needed. Musk took it a step further: in September 2017, he incorporated a public benefit entity called Open Artificial Intelligence Technologies, Inc. However, Musk demanded majority equity, absolute operational control, and the CEO title.

When OpenAI leadership rejected those terms, things fell apart quickly. By January 2018, Musk warned that OpenAI was on a path to certain failure unless it merged into Tesla. A month later, he stepped down as co-chair, later telling the team in December 2018 to raise billions per year immediately or forget it.

Suing versus building

OpenAI noted that this latest filing represents Musk's fourth attempt in less than a year to reframe his claims. In a direct punchline, the company wrote: You can’t sue your way to AGI.

The released emails also show early philosophical divides. Musk believed progress was moving too slowly, writing that linking deep nets didn't sound super hard and emphasizing the threat of Google's DeepMind. The research team pushed back, explaining that massive compute clusters help run bigger experiments 100 times faster, but raw hardware couldn't replace the key mathematical and conceptual breakthroughs they still lacked.

Why it matters

We think OpenAI's counteroffensive is exceptionally well-played from a PR perspective. By proving Musk registered his own for-profit vehicle and wanted unilateral control back in 2017, OpenAI effectively guts his public narrative of being a betrayed altruistic benefactor defending a non-profit mission.

For the broader industry, these emails confirm what insiders have long known: frontier AI research outgrew the charity model the moment neural networks began scaling with raw compute. Sutskever's 2017 estimate of a sub-$10 billion hardware bill looks quaint today, but it illustrates why a transition to commercial capital was inevitable for anyone serious about competing with Google.

That said, airing out dirty laundry carries risks. Laying bare the petty corporate power struggles, equity grabs, and ultimatum emails strips away the heroic sci-fi veneer from both Musk and OpenAI. It leaves observers with a much colder reality: this is simply a ruthless high-stakes tech rivalry disguised as an ideological war over humanity's future.