14,000 security workers at OpenAI, Anthropic authorize strike

Roughly 14,000 security officers guarding the California offices of OpenAI, Anthropic, Google, Meta, and Salesforce authorised an unfair-labour-practice strike on Wednesday, according to SEIU-United Service Workers West.

The workers aren't employed directly by the tech companies; they work for staffing agencies including Allied Universal, Securitas, and GardaWorld, which contract out building security across the Bay Area, Silicon Valley, and Los Angeles County.

Union officers marched through San Francisco and past the OpenAI and Anthropic offices to announce the vote. Whether an actual walkout follows, and when, remains unclear.

The vote caps roughly five months of negotiations that began around April 2026, with SEIU pushing Allied and the other agencies for higher wages, better healthcare, and more training.

"We're asking Allied and the other security companies to stop playing games, come to the table, and settle a strong contract," said Christopher Yates, an Allied Universal officer based in Hayward, in a statement distributed by the union.

The disagreement has already affected the operation of at least one AI firm even before the Wednesday vote. 

On Monday and Tuesday of last week, Anthropic instructed their staff in San Francisco to work remotely after Allied Universal warned of an impending strike by the company workers, despite SEIU’s claim at the time that it did not authorise nor threaten any strikes on those days.

The importance of workplace security in such firms is far greater than just another labour dispute. 

According to The Wall Street Journal, both Anthropic and OpenAI have been threatened against their workers, and it was claimed that in April, a Molotov cocktail was lobbed at the residence of OpenAI CEO Sam Altman. Altman later suggested that "doomerism talk" about AI may have contributed to the attack.