Two people were arrested after shots were fired near Sam Altman’s San Francisco home on Sunday, two days after a separate suspect threw a Molotov cocktail at the property and threatened to burn down OpenAI’s headquarters. The attacks coincided with a critical New Yorker profile of the OpenAI CEO, who responded with a blog post calling for de-escalation of anti-AI rhetoric.

Two people have been arrested after a gun was discharged near Sam Altman’s San Francisco home early on Sunday morning, the second security incident at the OpenAI chief executive’s Russian Hill property in three days.

The San Francisco Police Department confirmed the arrests of Amanda Tom, 25, and Muhamad Tarik Hussein, 23, both booked on charges of negligent discharge of a firearm. Officers responding to reports of gunfire at around 1:40am on Sunday traced a vehicle to a nearby address, where Tom and Hussein were detained without incident. Three firearms were seized from the property.

According to the SF Standard, a Honda sedan with two occupants stopped in front of Altman’s compound and the passenger appeared to fire a round on the Lombard Street side of the property. An OpenAI spokesperson, however, told Fox News Digital that the incident was unrelated to Altman and that there was no indication his home had been targeted. The SFPD has not publicly confirmed whether the shooting was directed at the property or coincidental.

Three days, two incidents

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

The shooting came 48 hours after a far more serious attack. In the early hours of Friday, 20-year-old Daniel Alejandro Moreno-Gama allegedly threw a Molotov cocktail at Altman’s home, setting fire to an external gate. No one was injured. Roughly an hour later, police arrested Moreno-Gama at OpenAI’s San Francisco headquarters, where he had allegedly gone to threaten to burn the building down.

Moreno-Gama was charged with attempted murder, arson of an inhabited structure, criminal threats, and possession of an incendiary device. His writings on Substack, dating back to January, expressed what investigators described as an existential fear that artificial intelligence would lead to human extinction. He had also been active in a Discord server run by PauseAI, a nonprofit that advocates for a temporary halt to frontier AI development. PauseAI issued a statement saying Moreno-Gama held no official role with the organisation and had been banned from the server after the attack. A moderator had previously warned him that advocating violence was grounds for removal.

The context Altman cannot escape

The attacks bookended the publication of a lengthy New Yorker investigation by Ronan Farrow and Andrew Marantz, based on interviews with more than 100 people, that examined Altman’s consolidation of power at OpenAI. Altman responded to both events in a Friday evening blog post that included a photograph of his husband and their child, a deliberate departure from his usual privacy, he said, intended to humanise the target.

“I underestimated the power of words and narratives,” he wrote, calling for a de-escalation of rhetoric around artificial intelligence. He acknowledged a pattern of conflict aversion that had “caused great pain for me and OpenAI,” referencing his removal and rapid reinstatement as chief executive in November 2023.

The property itself speaks to the scale of the tensions now surrounding AI leadership. Altman purchased the Russian Hill compound — spanning 855 Chestnut Street and the adjoining 952 and 954 Lombard Street,  in January 2025, through an affiliate managed by his cousin. The multi-lot acquisition, in one of San Francisco’s most visible neighbourhoods, has made him a conspicuous figure in a city where anti-technology sentiment has been intensifying.

The AI safety movement’s reckoning

The Molotov cocktail attack has forced an uncomfortable conversation within the AI safety community. PauseAI, which organises protests and lobbying campaigns calling for a moratorium on frontier model training, moved quickly to distance itself from Moreno-Gama. Its official statement noted that the organisation had previously warned him about violent rhetoric. But the incident has exposed a tension that safety advocates have long navigated: the line between urgently communicating existential risk and language that tips vulnerable individuals toward action.

Moreno-Gama’s Substack posts, which framed advanced AI as an imminent threat to human survival, drew on arguments that circulate widely in safety-focused communities, the difference being his apparent willingness to act on them. For organisations that have spent years arguing that AI development poses catastrophic risks, the attack creates a messaging problem that no press release can fully resolve.

A broader pattern

The incidents at Altman’s home are not occurring in isolation. A Fortune analysis published in December 2025 warned that Silicon Valley’s dismissive response to public anxiety about AI would have consequences in 2026, noting the widening gap between how the technology is framed by its builders and how it is experienced by the public. Public polling has consistently shown that a majority of Americans view AI’s trajectory with apprehension rather than optimism, and the rapid deployment of AI-generated content, automated hiring tools, and autonomous agents has made the technology’s impact tangible in ways that abstract safety debates never could.

The attacks also arrive at a moment of extraordinary flux for OpenAI itself. The company is in the process of converting from a nonprofit to a for-profit structure, a move that has drawn scrutiny from attorneys general in multiple US states. Its valuation has climbed past $300 billion. Altman, who controls no voting shares but wields outsized influence over the organisation’s direction, has become a lightning rod for anxieties that extend well beyond any single company.

The question of whether Sunday’s shooting was connected to Friday’s Molotov cocktail attack, or whether it was a coincidence of timing and geography,  remains formally unresolved. What is not in dispute is the sequence: a critical investigation, a firebombing, a plea for calm, and then gunfire, all within 72 hours. For the chief executive of the world’s most prominent AI company, the abstract debate about the technology’s risks has become acutely personal.

Tom and Hussein remain in custody. Moreno-Gama’s arraignment is pending. The SFPD says both investigations are ongoing.