Long-time OpenAI exec Brad Lightcap departs to 'start something new'

Brad Lightcap, one of the most visible leaders in OpenAI’s senior ranks, is heading out. He said on Tuesday that he’s leaving to “start something new,” though he didn’t say what that new thing is. He did say he’s “not going far,” and hinted that it will touch problems the world will need to get right as AI gets more capable.

That makes the move feel less like a clean break than a shift in seat. Lightcap joined OpenAI in 2018, before ChatGPT turned the company into a household name in late 2022. From there, he helped stand up the early versions of finance, legal, HR, corporate security, go-to-market, and partnerships — the boring machinery that lets a once-small lab act like a real company.

He was OpenAI’s CFO from 2018 to 2022, then became COO until April 2026, when he moved into a special-projects role while still reporting to Sam Altman. In his note on X, which he also sent to employees, he said he’d been thinking about “the next horizon” and about issues that could stand in the way of OpenAI’s AGI ambitions. He didn’t spell those issues out, but pointed to the kinds of challenges AI keeps running into as it moves deeper into public life.

The timing is awkward for OpenAI. The company is reportedly preparing for an IPO that could value it at more than $1 trillion, and it’s also under fresh scrutiny after recent reports that agentic models from OpenAI and Anthropic bypassed safeguards and hacked into other systems. Lightcap plans to stay for a few more weeks, and OpenAI has not publicly commented. Given his tenure and the amount of the company he helped assemble, this is not just another routine exit.

The broader pattern is simple: when a company starts acting like a platform for the whole industry, the people who built the plumbing don’t always want to spend forever inside the same machine. OpenAI’s talent churn is starting to look less like a blip and more like a feature of life at the center of AI.