OpenAI's Robotics Lead Quits Over Pentagon Deal

Caitlin Kalinowski, the person OpenAI hired to rebuild its robotics division from scratch, walked away from the company on March 7 — less than a year and a half after joining. Her reason: the Pentagon deal moved too fast on questions that should have moved slowly.

What Happened

Kalinowski's departure came days after OpenAI signed an agreement to deploy its models on a classified Pentagon network. The deal followed the collapse of negotiations between the Defense Department and Anthropic, which had refused to work without strict limits on domestic surveillance and autonomous weapons.

OpenAI stepped in where Anthropic wouldn't. For Kalinowski, that crossed a line.

"Surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got. This was about principle, not people." — Caitlin Kalinowski

Before OpenAI, Kalinowski spent years at Meta leading the creation of Orion AR glasses and nearly a decade at Oculus building VR headsets. She even helped design MacBooks during her time at Apple. OpenAI brought her on in November 2024 specifically to restart its robotics program, which the company had shuttered around 2021 to focus on language models.

Sam Altman acknowledged the deal's rollout looked "opportunistic." OpenAI later clarified that its agreement does include red lines — no domestic surveillance, no autonomous weapons — but critics note these are voluntary commitments, not enforceable restrictions. That distinction is exactly what Senate Democrats are now trying to fix through legislation.

Why This Matters

The resignation highlights a growing rift in the AI industry. On one side, companies like Anthropic insist on hard ethical boundaries even when it costs them government contracts. On the other, the pressure to compete for military dollars — and the geopolitical urgency of the US-China AI race — pushes companies toward flexibility.

For OpenAI's robotics ambitions specifically, losing Kalinowski is a real setback. She was the key hire for a division the company had only recently recommitted to. Recruiting top robotics talent may get harder when potential hires see that ethical objections lead to the exit door rather than the boardroom.

What's Next

Kalinowski said she remains "very focused on building responsible physical AI" but hasn't announced her next move. The broader debate about AI in national security is far from settled — with the Anthropic lawsuit still in court and multiple bills moving through Congress, the rules of engagement between Silicon Valley and the Pentagon are still being written.