The Atlantic says senior U.S. national-security officials accidentally included its editor in a Signal group chat about planned military strikes in Yemen, creating a rare and highly public glimpse into how sensitive government coordination can go wrong.
In the account, the magazine says it first received a connection request on Signal from a user identified as Michael Waltz on March 11, then was added two days later to a group called the “Houthi PC small group.” According to the article, the chat included officials identified as senior figures across the administration and was used to discuss upcoming action against Houthi targets. The publication says the first notice of an attack came before the bombing began, allowing it to infer that the chat was real rather than a hoax.
The piece says Pete Hegseth texted the war plan at 11:44 a.m. and that the attack began shortly after 2 p.m. eastern time on March 15. It also says the message thread included operational details such as weapons packages, targets, timing and follow-up coordination. The magazine does not describe the contents as a general policy discussion. Instead, it presents the thread as evidence that an apparently classified or highly sensitive conversation was being conducted on a commercial messaging app.
The story also lays out how the group formed. The Atlantic says a message from Michael Waltz described a “principles” group for coordination on the Houthis and asked each department to provide a staff point of contact. Later messages in the article show users identified as Marco Rubio, JD Vance, Tulsi Gabbard, Scott Bessent, Pete Hegseth, Brian McCormack, John Ratcliffe and others assigning representatives or commenting on the plan. The article says the group eventually reached 18 members.
Beyond the leak itself, the article underscores the tension between convenience and secrecy. Signal is widely used because it offers encrypted communication that feels more private than standard texting. But The Atlantic’s account argues that the same convenience can create the risk of accidental exposure if senior officials use it for discussions that should stay on secure government systems. The publication says it struggled to believe the chat was genuine because it seemed implausible that such a conversation would happen on a consumer app, yet the timing of the strikes confirmed that it had.
The article also captures a policy disagreement inside the group. It quotes a message attributed to JD Vance saying the administration was “making a mistake,” with concerns about how the public might understand the operation and why it was necessary. That detail matters because it shows the chat was not simply a logistics feed. The messages, as reported, included strategic debate alongside tactical planning.
For national security, the episode is significant not because it reveals a long-term policy shift, but because it exposes a breakdown in process. The article suggests that the failure was procedural and operational: a conversation involving imminent military action was taking place on a consumer platform, with a journalist present in the thread. If accurate, that would represent a major communications lapse inside one of the most sensitive parts of government.
The Atlantic says it consulted colleagues because it initially suspected a disinformation attempt. Its eventual publication turns the incident into a cautionary story about how quickly secure and insecure channels can blur when high-level officials rely on ordinary messaging tools.
