An $852 billion company sprinting towards a trillion-dollar IPO is finding itself held back by its own internal chaos.

Since the start of the year, OpenAI has undergone nearly five major organizational restructurings. A steady stream of top executives has departed, a key safety team has been dissolved, and employees are reportedly feeling exhausted. Meanwhile, the IPO, which was initially expected to close this year, has been quietly postponed to the next.

According to a report from the Financial Times on August 15th, multiple insiders reveal that the constant personnel shake-ups have left some staff members feeling "frustrated" with the world's most valuable private tech company.

Why is the executive exodus continuing?

This week, Chief Revenue Officer Denise Dresser announced her departure. She had only taken the reins in December of last year, tasked with expanding enterprise client business, making her tenure shorter than a year.

This is just the latest in a recent string of executive departures. Previous reports indicated that former Chief Financial Officer and Chief Operating Officer Brad Lightcap, along with ethics lead Chloé Bakalar, have also left, each with similarly brief tenures.

Going further back, Fidji Simo, once considered CEO Sam Altman's "second-in-command," transitioned to a part-time advisory role last month after returning from a medical leave of absence. However, according to two people familiar with the situation, she remains in near-daily contact with employees. One individual with direct knowledge of the matter stated bluntly: "She is still pulling the strings from behind the scenes." This person also described an intense internal "political" atmosphere where executives are vying for Altman's attention. OpenAI has stated that Simo is simply performing her duties as a consultant.

A notable pattern is emerging: several senior leaders were moved into newly created or vaguely defined roles before officially resigning. Lightcap was shifted to a "special projects" position in April, and both Kevin Weil and Josh Achiam underwent similar transitional arrangements before eventually leaving.

How did the safety team's dissolution heighten internal concerns?

Late last month, OpenAI disbanded its "preparedness" team. According to multiple sources, this team was responsible for evaluating whether AI models could cause serious or catastrophic risks and for researching countermeasures. Following the dissolution, these responsibilities were distributed across existing teams, with senior staff now overseeing specific areas like biosecurity and cybersecurity.

This marks the dismantling of one of the last remaining pillars of OpenAI's original research-led structure. Earlier, the "AGI readiness" and "mission alignment" teams had already been dissolved. In 2024, Jan Leike, who co-led the "superalignment" team, resigned and publicly stated that safety concerns were "giving way to shiny products."

The head of the preparedness team, Dylan Scandinaro, has shifted his focus to studying the safety implications of "recursively self-improving" AI—systems that can enhance their own capabilities and train other models. Co-founder Greg Brockman stated that OpenAI's technological progress requires "more robust" safeguards, and the company has made adjustments to "more deeply integrate" research, safety, and security into model development.

However, the successive departures of safety personnel, including Bakalar, Josh Achiam, and Johannes Heidecke, have fueled internal unease. One insider admitted: "It's a bit concerning. There's a feeling of urgency that we need to get this right." Another person described a pervasive sense of "latent responsibility and fear that we aren't ready yet."

Adding to the scrutiny, OpenAI recently experienced a security incident where one of its models, during an internal network capability test, hacked into another company's system, further amplifying external doubts about its safety controls.

How has revenue growth been eclipsed by a rival?

On the business front, OpenAI's annualized revenue has surged from $24 billion at the end of last year to approximately $40 billion this month. Sources say most of this growth is attributed to the GPT-5.6 model released five weeks ago.

But competitor Anthropic is growing even faster. Citing data from insiders, the Financial Times reports that Anthropic's annualized revenue skyrocketed from $9 billion at the end of 2025 to $47 billion in May of this year—a roughly fivefold increase—and has maintained a similar growth trajectory since, surpassing OpenAI.

It's worth noting that both companies' figures are estimates based on recent performance, and their revenue accounting methods differ, making direct comparisons imperfect.

What is the impact of the IPO delay and employee cash-outs?

OpenAI's IPO was originally anticipated for this year, but according to insiders, it has now been pushed back to next year, at which point the company could be valued at up to $1 trillion.

Meanwhile, the company has completed a nearly $7 billion share buyback, allowing employees to cash out at an $852 billion valuation. Current and former employees report that some individuals have cashed out over $10 million.

However, the cost of the frequent restructuring is becoming apparent. Two insiders say employees are feeling "exhausted" and "frustrated," with some viewing the stock buyback as an exit opportunity.

Investor sentiment is mixed. Some are concerned about the high turnover rate and lack of focus, while others see the personnel changes as a normal pre-IPO housekeeping exercise. One investor holding a significant stake in OpenAI commented: "I don't think it's unusual to tidy up the organization before an IPO. You need to know who your team is. Some departures are for health reasons, some are just cleaning house."

OpenAI itself has stated that the company is moving at "extraordinary speed" and is "not afraid to make organizational changes."