The Donation That Wasn't: OpenAI's Greg Brockman Walks Back a $25M Pledge as AI Politics Boils

Greg Brockman, OpenAI’s co-founder and president, has backed out of a planned second $25 million donation to Leading the Future, the pro-AI super PAC he helped bankroll, according to a report published by The New York Times on September 30. In an internal message to staff, Brockman and his wife Anna said the group had become a “distraction,” walking away from half of the $50 million the couple had originally intended to pour into the 2026 election cycle.

What Happened

The reversal, confirmed by The Wall Street Journal, caps months of escalating friction inside OpenAI over the political turn its leadership has taken. Last September, the Brockmans donated an initial $25 million to Leading the Future, a bipartisan super PAC that has since committed more than $140 million — with backing from figures tied to OpenAI, Palantir, and Andreessen Horowitz — to shape state and federal AI policy debates. The couple had pledged a matching second tranche for 2026. That pledge is now dead.

Brockman’s June Slack message to employees, acknowledging the turmoil the first donation created, foreshadowed the retreat. According to the Journal, he told staff this week that he would not follow through on the remaining $25 million, framing the decision as an attempt to defuse internal anger rather than a change of conviction. He has previously insisted that his donations — including another $25 million he and his wife gave last year to Donald Trump’s super PAC — were made in a purely personal capacity, with no coordination from OpenAI itself.

The Backlash Was Already Boiling

To understand why a billionaire would walk away from $25 million of political intent, look no further than his own workforce.

OpenAI has spent the summer at war with itself over exactly this money. In July, Wired reported that rank-and-file OpenAI employees had donated more than $215,000 to the Guardrails Alliance, a rival tech-worker-backed super PAC whose explicit purpose is to counter Leading the Future. Eight contributions from current and former OpenAI staff eventually totaled roughly $248,000 — pocket change next to $140 million, but unmistakable as a signal. Two of the donors, Julie Steele and Jason Wolfe, work on AI alignment research at the company: the people whose job is to make OpenAI’s models safe, funding a campaign against their own president’s political project.

The Guardrails Alliance, launched in June with about $5 million raised and positioning itself as a populist counterweight to pro-AI money, has made “you can’t out-spend $140 million, you can out-organize it” its rallying cry. Its existence — and the fact that OpenAI’s own alignment researchers were among its earliest donors — turned Brockman’s checkbook into a running internal controversy.

OpenAI the company has, for its part, repeatedly distanced itself from Brockman’s giving. A June statement reported by Business Insider said the company had nothing to do with the donations. Transformer reported in June on tense internal meetings where employees pressed leadership over the gap between the company’s stated positions and the political spending of its co-founder. The New York Times separately documented how Leading the Future’s advertising has coursed through races where AI regulation itself is on the ballot — including attack ads in New York’s legislative fights, where Governor Hochul has signed AI bills that LTF-aligned money opposed.

Why It Matters

Three things make this more than a personal finance story.

First, it marks the first high-profile defection from the big-money pro-AI bloc. The 2026 midterms have become the most expensive AI-policy fight in American history, with industry PACs, safety groups, and tech-worker organizations pouring tens of millions into races hinging on state AI laws. Leading the Future’s $140 million war chest was the single largest flag planted. Brockman’s withdrawal doesn’t cripple the group, but it cracks the narrative of an invincible, unified industry front — and hands a rhetorical victory to the safety side, which had spent months arguing that the money was resented even inside the companies it came from.

Second, it exposes how deep the internal rift at OpenAI runs. This is a company whose president just publicly retreated from a political position because his own employees made it untenable. That dynamic — alignment researchers funding opposition PACs, all-hands meetings consumed by donation controversy, a co-founder conceding “distraction” — would have been unthinkable at any major tech firm of the last decade. It suggests political spending by AI executives is now a first-order internal-relations liability, not just a reputational one.

Third, the timing is brutal. The reversal lands days after the White House “super intelligence” accord signed by executives from Google, Anthropic, Meta, Nvidia, OpenAI, and xAI — an event that already put the industry’s Washington posture under a microscope — and in the same week that a Senate hearing on rogue AI left the industry facing pointed questions. With the November elections five weeks away, every dollar moving in or out of these PACs is now front-page news.

What Comes Next

Brockman’s first $25 million remains with Leading the Future — the withdrawal only kills the second tranche — and the PAC’s other funders, including major venture figures, have not followed suit. The Guardrails Alliance, meanwhile, continues to organize tech workers on a shoestring budget relative to its adversary.

But the signal is sent: the AI industry’s political consolidation has limits, and those limits are being enforced from the inside. For a sector that has spent 2026 learning to throw its weight around in Washington, this week offered the first vivid demonstration of the weight swinging back.

Sources are listed in the article metadata.