IN June, at Joint Base Myer–Henderson Hall in Arlington, Virginia, four current and former executives from Meta, OpenAI and Palantir stood on stage and raised their right hands.
In front of a military audience, they swore an oath to support and defend the United States.
It wasn’t a cameo.
The US Army had created a new technical innovation detachment, and the executives were being formally inducted as lieutenant-colonels in the unit, known as Detachment 201.
Their job: advise on emerging technologies that could change the face of combat.
“We desperately need what they are good at,” Secretary of the Army Daniel Driscoll said. “It’s an understatement how grateful we are that they are taking this risk to come and try to build this out with us.”
The executives have since gone through basic training.
That ceremony was no one-off. In the Trump era, the US military hasn’t merely courted Silicon Valley. It has successfully recruited it.
Over the past two years, leaders who once insisted their companies should steer clear of weaponry have plunged headfirst into the defence sector.
Meta, Google and OpenAI – all of which previously had policies explicitly banning the use of artificial intelligence for weapons – have quietly removed that language.
Defence-related start-ups have surged and venture capital investment in the sector jumped 33% last year to US$31bil, according to McKinsey.
A cultural shift is under way, reshaping a sector once squeamish about the Pentagon.
“A decade ago, people were talking about ‘connecting the world’ and ‘do no evil’,” Andrew Bosworth, Meta’s chief technology officer and one of Detachment 201’s new lieutenant-colonels, said at a tech conference in San Francisco.
“Now the tides have turned... There’s a much stronger patriotic underpinning than I think people give Silicon Valley credit for.”
The driving forces
Several pressures have combined to push Silicon Valley into the military’s embrace.
The mounting rivalry with China over technological leadership has concentrated minds in Washington and California alike.
The wars in Ukraine and Gaza have shown how decisive drones and AI- assisted weapons have become.
At the same time, Trump’s domestic policy package earmarks a record defence budget for 2026, including major spending on autonomous systems.
For venture capitalists, that is a bonanza. For some engineers, it is a moral minefield.
Once they build autonomous drones or AI models for weapons systems, they lose control over how the technology is used.
Three engineers at Google and Meta said the debate inside their companies has become increasingly fraught: will these systems save lives on the battlefield, or take more?
“These Silicon Valley companies are hypercompetitive, and in their drive to get into these defence sectors, there isn’t a lot of pausing to think,” said Margaret O’Mara, a University of Washington historian who studies the tech industry.
A return to roots
In many ways, this shift isn’t a break from Silicon Valley’s past but a return to it.
In the 1950s, the Department of Defence poured money into early computing and electronics firms to counter the Soviet Union.
The Pentagon was one of the region’s earliest backers.
That long history faded from view during the 2010s, when working with the US government became a flashpoint.
In 2018, more than 4,000 Google employees protested against Project Maven, a Pentagon initiative to use Google’s AI to analyse drone footage.
Workers declared that Google “should not be in the business of war”.
The company backed away from the contract and withdrew from bidding for a separate US$10bil cloud deal, known as Jedi.
Google then published AI principles forbidding work on “weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people”. Other firms followed suit.
There were exceptions
Palantir’s chief executive, Alex Karp, pushed aggressively for deeper tech–military collaboration.
He sued the army in 2016 to force it to evaluate Palantir’s software – and won.
The ruling set a precedent for other defence-focused start-ups.
Proud to engage
Now the pendulum has swung.
In January 2024, OpenAI removed wording that prohibited using its models for “weapons development” and “military and warfare”.
Months later, it announced a partnership with Anduril, a fast-rising defence firm, to build AI systems for counter- drone operations.
“We have to, and are proud to, and really want to engage in national security areas,” OpenAI chief executive Sam Altman said.
The company would help develop AI “supporting the US and our allies to uphold democratic values around the world and keep us safe”.
Meta followed suit, revising its policies last year to allow its AI to be used in military settings.
In May, it announced a collaboration with Anduril to build virtual-reality training tools for soldiers.
Google made the same turn in February. In a blog post, the company declared: “We believe democracies should lead in AI development.”
The Defence Department’s enlistment of the four tech executives – Bosworth, Palantir CTO Shyam Sankar, OpenAI chief product officer Kevin Weil and Bob McGrew, an adviser at Thinking Machines Lab and formerly OpenAI’s chief research officer – signals just how far the Valley’s mindset has shifted. — ©2025 The New York Times Company
This article originally appeared in The New York Times
