Q&A with Palantir CTO Shyam Sankar on surveillance, AI, the future of war, data, privacy, the US military, ImmigrationOS, the DHS, ICE, the UK's IDs, and more
The tech company's C.T.O. on surveillance, A.I. and the future of war. — Few companies in Silicon Valley inspire as much paranoia as Palantir. X: @douthatnyt and @morgangress Bluesky: @drewharwell.com and @justinhendrix X: Ross Douthat / @douthatnyt : This week's Interesting Times is a conversation with Shyam Sankar, the CTO of Palantir and a lieutenant colonel in the US Army Reserves: https://www.nytimes.com/... Morgan Gress / @morgangress : So good, a real conversation about what @PalantirTech actually does — simplified and grounded in the tech and mission. @ssankar & @DouthatNYT don't shy away from the complex issues either (no surprise there!)👏https://www.nytimes.com / ... Bluesky: Drew Harwell / @drewharwell.com : My 'We're not building Skynet' shirt has people asking a lot of questions already answered by my shirt [embedded post] Justin Hendrix / @justinhendrix : There is a long interview with Palantir CTO Shyam Sankar by columnist Ross Douthat in the Times today. “We're explicitly not building Skynet,” he reassures Douthat, in a wide ranging conversation that ranges from the company's work with ICE and Israel to Sankar's appetite for pro-American movies.
Context & Ripple Effects
Palantir’s public case for its role in security and defense has been building alongside scrutiny of its government work. Earlier coverage documented its deep integration into Ukraine’s wartime government operations and a company-backed forum that showcased AI-enabled military tools, including systems for identifying potential targets.
This interview puts its CTO—and U.S. Army Reserve officer—at the center of that debate. It matters less as a product announcement than as an attempt to define the limits of Palantir’s surveillance and autonomous-AI posture amid enduring questions about the company’s trustworthiness.
First-order effects
- Palantir gains a high-profile venue to contest the idea that its systems amount to unconstrained autonomous surveillance or an all-powerful AI, while explaining its work across military, DHS and ICE-linked contexts.
- Customers, critics and policymakers get a more explicit public account of how a senior Palantir executive frames the boundary between decision-support software and autonomous action.
Second-order effects
- The interview increases pressure on other defense- and government-tech vendors to articulate comparable safeguards and human-accountability claims when selling AI into sensitive public-sector workflows.
- Debate over Palantir’s ICE and DHS relationships is likely to remain tied to procurement legitimacy: technical capability alone will not settle concerns about how government users deploy it.
Third-order effects
- If defense and domestic-security AI becomes normalized through operational deployments and public justification, vendors’ political legitimacy and governance claims will become a durable competitive factor alongside software performance.
- The central fault line is likely to be whether AI systems augment accountable human decisions or widen institutions’ capacity for surveillance and targeting; that distinction will shape public acceptance and oversight.
The trend: Palantir is part of a broader shift in which AI vendors seek strategic legitimacy by tying advanced software to national-security missions while defending its use against surveillance and autonomy concerns.