This week’s defence AI landscape is defined by three converging pressures: the hardening of AI into operational warfighting infrastructure (from Palantir’s Army data overhaul to the Booz Allen edge-AI doctrine), a deepening regulatory fault line between the EU’s structured governance push and the UK’s more laissez-faire approach, and alarming new evidence that AI deepfakes are now being deployed as a standard weapon of war — not merely a propaganda nuisance. Across NATO allies, the race to operationalise AI faster than adversaries is visibly outrunning the governance frameworks designed to keep it accountable.
Top story: Palantir’s Foundry platform is chosen as the core cloud data layer for the US Army’s highest-priority modernisation programme, NGC2, cementing it as the backbone of American military AI infrastructure.
Palantir Chosen as Core Data Layer for US Army’s Top AI Overhaul
Timothy Sykes / PLTR News · Strategy
The US Army selected Palantir’s Foundry as the central cloud data layer for its Next-Generation Command and Control (NGC2) programme, its highest-priority modernisation effort, as the platform shifts from prototype to wide deployment. A separate Palantir–Nvidia initiative will run Nvidia AI and Nemotron models in secure, sovereign environments for sensitive government and critical infrastructure workloads. Together these moves signal Palantir is transitioning from contractor to essential infrastructure for US defence AI.
https://www.timothysykes.com/news/palantir-technologies-inc-pltr-news-2026_07_02/
AI Deepfakes Weaponised in Combat: A New Doctrine Emerges
Bloomsbury Intelligence and Security Institute · Risk
A new analysis from the UK’s Bloomsbury Intelligence and Security Institute documents how AI deepfakes were integrated with kinetic military operations during the Iran-Israel conflict, with AI-generated videos of fabricated missile strikes spreading across platforms in five languages within hours of real attacks. Crucially, Israel’s own Operation PRISONBREAK deployed deepfake content within one hour of actual strikes, showing that even democratic states now treat AI-enabled psychological operations as essential military capabilities. The report warns that AI-powered disinformation will likely achieve decisive advantage over detection capabilities by late 2026.
Helsing and OHB Launch AI Space-Based Targeting System ‘Kirk’
Business Standard / Blueprint Defence · Generative AI
German defence AI firm Helsing and aerospace company OHB have launched the ‘Kirk’ initiative to develop AI-enabled, space-based targeting and intelligence capabilities, integrating satellite imagery, AI, and real-time data processing to rapidly scan and engage military targets. Designed explicitly for European defence requirements, the system aims to shorten battlefield decision-making timelines and improve situational awareness. The move deepens European sovereign AI defence capacity at a moment when the continent is racing to reduce reliance on US-controlled platforms.
Booz Allen: Military AI Must Reach Space and Every Soldier’s Pocket
National Defense Magazine · Strategy
A high-profile Booz Allen analysis published this week argues that winning modern conflicts requires AI to operate simultaneously at the strategic edge — from satellites to exo-atmospheric interceptors — and at the individual operator level, all without guaranteed connectivity. The piece makes the case for ‘edge-first AI architectures’ that can orchestrate in highly unpredictable network environments, enabling split-second survivability and lethality decisions. It signals a major shift in Pentagon thinking: AI is no longer a back-office tool but must be embedded in every layer of the kill chain.
https://www.nationaldefensemagazine.org/articles/2026/7/21/sponsored-content-booz-allen-july-2026