OpenAI executive exodus deepens ahead of possible $1 trillion IPO

OpenAI is facing growing internal turmoil after roughly five organizational changes since the start of the year, as CEO Sam Altman races against time to prepare the company for an IPO that could become the largest in tech history, according to people familiar with the matter cited by the Financial Times. The report was independently confirmed by The Verge and Engadget, while the Arabic-language tech outlet aitnews cited the British newspaper.

Senior leaders head for the exit

Denise Dresser, chief revenue officer, announced her departure this week, less than a year after being appointed in December to court enterprise customers. Her exit follows those of Brad Lightcap, the former chief financial and operating officer, and Chloe Bakalar, who oversaw ethics, both of whom left after less than a year in their roles. Fidji Simo, who ran the applications unit and was seen as the most influential leader after Altman, also stepped down from her executive role last month following a medical leave, moving into an advisory position — though according to the newspaper she remains in close contact with staff and involved in some internal matters, amid competition among senior executives for Altman's attention.

The AI risk-assessment team is broken up

OpenAI saw another shake-up last month with the disbanding of its 'Preparedness' team, which assessed the likelihood that the company's models could cause severe or catastrophic harm and developed measures to limit that risk. Rather than keeping it as a standalone unit, its responsibilities were distributed across different teams, with specific leaders assigned to areas such as cybersecurity and biological risk. The move is part of a broader push to simplify the company's structure, after Altman asked staff to cut back on what he called 'side quests' and focus on ChatGPT and competing with Anthropic for enterprise customers. Since the governance crisis that briefly ousted Altman in late 2023, he has reshaped his leadership team to make the company's operations more professional, while co-founder Greg Brockman has gained growing influence as several leadership positions remain vacant. Brockman said the company's technical progress requires 'more robust' safeguards, adding that the changes are meant to integrate research, safety and security more deeply into model development.

A near-trillion-dollar valuation amid a fierce race with Anthropic

Some employees see the pace of reorganization as reflecting administrative disarray at a sensitive moment, as OpenAI prepares for an IPO that could push its valuation to nearly $1 trillion. The company had been expected to go public this year, but the move now looks closer to 2027, according to people familiar with its plans. OpenAI faces mounting pressure from rival Anthropic: its own estimated annual revenue rose from about $24 billion at the end of last year to nearly $40 billion this month, yet Anthropic has overtaken it after its estimated annual revenue jumped from $9 billion at the end of 2025 to $47 billion in May, continuing to grow at a similar pace since.

• 'Preparedness' catastrophic-risk team: disbanded last month, its work spread across other teams
• OpenAI's estimated annual revenue: from $24 billion at the end of last year to about $40 billion this month
• Anthropic's estimated annual revenue: from $9 billion at the end of 2025 to $47 billion in May
• Expected IPO: a valuation that could approach $1 trillion, likely pushed back to 2027

What it means for businesses betting on OpenAI's tools

For businesses embedding ChatGPT or the OpenAI API in their products or internal operations, this pattern of leadership churn and the dismantling of dedicated safety and oversight structures is worth tracking separately from the IPO story itself: it suggests that the people and internal processes determining how OpenAI assesses risk before shipping new capabilities are less stable than the outwardly steady cadence of product releases would suggest. Companies with regulatory obligations, particularly in sectors like finance and healthcare, should treat OpenAI's internal risk review as a moving target and maintain their own risk-assessment process for each model update, rather than assuming continuity in how the company evaluates safety internally.