- Japan is considering using U.S. AI systems like Palantir and Anduril to speed up its military decision-making.
- Critics warn that giving foreign corporations access to classified Japanese data risks national sovereignty.
- Palantir’s Maven Smart System was only 60% accurate in target selection, well below human analysts.
- A 2025 UN report named Palantir among companies sustaining an economy of genocide over its Gaza AI tools.
- Japan plans to build domestic oversight centers but admits reliance on U.S. AI is unavoidable for now.
Japan is considering integrating U.S.-built artificial intelligence platforms, including Palantir Technologies' Maven Smart System and Anduril Industries' Lattice, into the command-and-control structure of its Self-Defense Force, according to a Nikkei Asia report published August 5. The move aims to close a growing decision-making gap with American forces, but critics warn it would give foreign corporations access to classified Japanese military data.
Why the Pentagon's AI is coming to Tokyo
Japanese military commanders currently operate without AI assistance, while U.S. forces have deployed these systems for years. That disparity threatens to disrupt joint operations between the two allies. The SDF established a Joint Operations Command in 2025 and has deepened cooperation with American troops stationed in Japan and across the Pacific, but slower Japanese response times risk breaking the chain of command in combat situations, according to Nikkei Asia.
Tokyo envisions deploying separate AI platforms for counterattacks, missile defense and drone control. But these systems would require access to classified SDF material, raising concerns about foreign corporate oversight and potential limits on services and functions. Tokyo hopes to mitigate the risk by keeping open the option to replace the U.S. technology with domestic AI later, and by building data centers and other infrastructure within Japan.
A vendor with a track record worth scrutinizing
Handing a foreign contractor the keys to a nation's military decision-making would be a hard sell under any circumstances. It's a harder one with Palantir specifically. Co-founded with backing from the CIA's venture capital arm In-Q-Tel, the company has drawn blacklisting and contract cancellations across U.S. allies in recent months. Spain barred Palantir from public and state-controlled entities in July over security concerns, and governments in France, Germany and the United Kingdom have since moved away from the firm as well, citing similar worries about handing sensitive state data to a private U.S. company.
Much of that scrutiny traces back to Gaza. UN special rapporteur Francesca Albanese named Palantir in a 2025 report among 48 companies she said are sustaining what she termed an "economy of genocide," pointing to AI tools the company provided Israel after October 2023 that Israeli forces have used to help generate strike lists.
Palantir's other flagship product hasn't inspired much more confidence. The Maven Smart System, used extensively by U.S. forces in the war on Iran earlier this year, reportedly produced over 3,000 proposed targets in a single day — but got them right only about 60 percent of the time, well below the 84 percent accuracy of human analysts.
By April 9, U.S. Central Command had recorded a barrage of more than 13,000 strikes across Iran. One of them hit the Shajareh Tayyebeh Primary School in Minab, killing 168 people — roughly 120 of them students between seven and twelve years old. Iran's IRGC responded by naming Palantir among 18 U.S. tech companies whose regional facilities it now considers fair game for retaliation.
Palantir has also come under fire for booking a global effective tax rate of just 1.4% last year, including $0 in US federal income tax, despite more than $1.6 billion in pre-tax profit — a track record critics say undercuts the company's pitch as a trustworthy steward of public money and public trust.
Sovereignty on paper, dependence in practice
Japanese officials acknowledge the need to balance speed with sovereignty. While there is "a strong desire for AI sovereignty through the use of domestic models in Japan's defense," the government treats reliance on U.S. systems as "unavoidable" if capabilities are to be raised quickly, Nikkei reported.
One proposal would deploy multiple U.S. AI systems overseen by a Japanese-made "command center" that decides which platform to use in each situation. A model developed by Tokyo-based Sakana AI — which already holds a Defense Ministry contract — is a strong contender for that oversight role, although it's unclear whether the American and Japanese systems would work well together, or whether the arrangement would simply entrench the U.S. technology further.
Japan's dilemma is a familiar one for U.S. allies: buy speed now from a company already blacklisted by half of Europe and accused of enabling mass civilian casualties abroad, or build something homegrown and fall behind. Tokyo appears set to choose the former — and to hope its hedges are enough.
Sources for this article include:
Please contact us for more information.
