OpenAI under strain

OpenAI spent Friday answering three very different problems at once: a security warning for its Mac apps, a courtroom fight with Elon Musk, and a police investigation after a Molotov cocktail was thrown at Sam Altman’s San Francisco home. The mix is unusual because one problem hits product trust, one hits corporate control, and one hits basic executive security. (openai.com) (bloomberg.com) (japantimes.co.jp) The product issue started with a third-party developer tool called Axios, which OpenAI said was part of a broader industry incident. OpenAI said it found no evidence that user data was accessed, its systems were breached, its software was altered, or its intellectual property was compromised. (openai.com) (indianexpress.com) The weak point was the process that tells a Mac computer an app is really from OpenAI. OpenAI said it is updating those security certifications and requiring Mac users to install the latest versions of its apps so an attacker cannot try to pass off a fake desktop app as a real one. (openai.com) (cnbc.com) That kind of warning lands differently for an artificial intelligence company than for a game studio or a shopping app. OpenAI’s desktop products can sit next to source code, documents, and internal chat windows, so even a certification scare raises the question of what users are willing to trust on their laptops. (openai.com) (cnbc.com) At the same time, OpenAI is preparing for a huge trial with Elon Musk, who helped found the company in 2015 and later broke with Sam Altman over its direction. Musk sued OpenAI and Microsoft in 2024, arguing that the company abandoned its original nonprofit-style mission when it built a for-profit structure around ChatGPT. (moneycontrol.com) (bloomberg.com) OpenAI told the court this week that Musk changed what he is asking for just weeks before trial. Bloomberg reported that Musk’s new proposals include sending any trial winnings back to OpenAI, unwinding OpenAI’s conversion, and giving oversight over financings and transactions, which OpenAI called a legal ambush that would require different evidence and witnesses. (bloomberg.com) (business-standard.com) The dollar figure hanging over that case is enormous because OpenAI is no longer a lab project with a few donors. Bloomberg described the coming case as a $100 billion-plus trial, which means the argument is not just about old emails and founding ideals but about who gets to shape one of the most valuable private technology companies in the world. (bloomberg.com) Then the story turned physical. San Francisco police arrested a man after a Molotov cocktail was thrown at Altman’s home early Friday, and Reuters reported the same suspect then threatened to burn down OpenAI’s headquarters in San Francisco. No one was injured. (cnbc.com) (aljazeera.com) Those three events are connected by timing more than by cause, but timing matters when a company sells trust. On April 10 and April 11, 2026, OpenAI had to tell users its Mac app chain needed urgent fixes, tell a judge Musk had changed his demands, and tell the public its chief executive had been targeted at home. (openai.com) (bloomberg.com) (cnbc.com) None of this means ChatGPT suddenly stopped working or that OpenAI said customer data was stolen. It means the company that wants its tools inside schools, offices, phones, and operating systems is being tested on the less glamorous parts of power: software signing, courtroom discipline, and personal security. (openai.com) (bloomberg.com)