14,000 Security Guards Plan to Strike at OpenAI, Anthropic and Other Silicon Valley AI Firms

A five-month contract fight between security contractors and the guards protecting OpenAI, Anthropic, Google, Meta and Salesforce is set to boil over Friday, against a backdrop of real threats against AI executives and offices.

Roughly 14,000 security officers who guard the offices of OpenAI, Anthropic, Google, Meta and Salesforce are set to walk off the job Friday, according to the union representing them (KQED).

It’s an unfair labor practice strike, not a garden-variety contract dispute. It caps five months of bargaining that’s gone about as smoothly as you’d expect when one side is asking for $30 an hour and the other keeps countering with pocket change.

Guarded by Contractors, Not the Companies Themselves

The workers aren’t employed by the tech companies whose lobbies and campuses they staff. They’re hired through third-party contractors, mainly Allied Universal and Securitas, along with smaller firms like PalAmerican and GardaWorld.

That arrangement is exactly why a security guard standing outside a company minting AI models worth hundreds of billions of dollars can end up making $20.09 an hour, the current median wage per a UC Berkeley study cited by KQED. The AI firms don’t sign these paychecks. The contractors do, and the contractors have been slow to move.

The Numbers Behind the Standoff

Service Employees International Union-United Service Workers West, which represents the officers, says the latest offer on the table is a 65-cent raise spread over two years.

Earlier in the negotiations, the number was even worse: 25 cents an hour in 2027, with nothing after that for three more years. SEIU-USWW president David Huerta called that proposal “shameful, wrong, and deeply, deeply disrespectful”, which is union-speak for “you’ve got to be kidding me.”

The union frames the dispute as more than a pay gap. Its strike announcement cites bad faith bargaining, surveillance of officers, retaliatory discipline and interference with workers’ right to organize as additional grounds for the walkout.

Allied Universal has said it doesn’t comment on active negotiations. Securitas didn’t respond to requests for comment either, which is its own kind of comment.

A Face on the Numbers

The human details are doing a lot of the persuading here.

Jerry Longoria, an Allied Universal officer who guards a Cartier store and has sat through five contract negotiations in 25 years on the job, told KQED he lives in an SRO without his own bathroom or kitchen.

He’s watched a customer buy a $1.5 million necklace at the counter he protects (48 Hills). Longoria isn’t asking for a cut of that. He’s asking for rent money.

Why This Isn’t Just a Labor Story

There’s a part of this dispute that makes it more than a wage negotiation: AI firms’ security spending has spiked because their executives and offices have become actual targets, not hypothetical ones.

In April, a 20-year-old man threw a Molotov cocktail at OpenAI CEO Sam Altman’s San Francisco home before dawn, setting a gate on fire, then walked to OpenAI’s headquarters and threatened to burn it down. He was arrested and later charged with attempted murder, including a second count for the security guard on scene at Altman’s home (NPR).

Days later, gunfire hit Altman’s house. Altman later wrote that he’d “underestimated the power of words and narratives” and called for the industry to de-escalate its rhetoric, literal explosions included.

The Wall Street Journal has separately reported that San Francisco police have fielded a number of threats against OpenAI and Anthropic employees more broadly. That part is well documented.

What isn’t confirmed is any specific new threat tied to this week’s strike vote itself. The connection reporters are drawing is broader: AI firms now depend more than ever on the guards they’re currently underpaying.

A Preview That Backfired

That dependency showed up in an odd dry run earlier in September. Anthropic told its San Francisco staff to work from home for two days after Allied Universal warned of a possible walkout by its guards (TechRadar).

SEIU says it hadn’t authorized or even threatened a strike on those specific days, which left the union about as confused by Anthropic’s move as everyone else. Call it a false alarm, or call it a company getting nervous about what happens when the people watching the door decide to stop.

What Happens Friday

The guards marched through downtown San Francisco this week, past Salesforce’s Dreamforce conference, where Sam Altman, Dario Amodei and Marc Benioff were all expected to appear.

Whether Friday’s walkout actually disrupts operations at any of these companies remains to be seen. But the optics are hard to miss: an industry racing toward AGI, still leaning on humans with badges to keep the front door secure, and currently unwilling to pay those humans enough to live near the buildings they guard.

Further reading: Business Insider, New York Post, KRON4