This week, AI in defence moved decisively from strategy to procurement and infrastructure, with the UK’s landmark £2bn AI Battle Lab contract the clearest signal yet that Western militaries are embedding AI into core training and readiness systems at scale. Simultaneously, the US White House’s NSPM-11 directive is tightening vendor accountability for all national security AI, while Australia’s live deployment of a German AI surveillance drone underscores how allied forces are already fielding AI-powered ISR in joint exercises. The gap between nations with funded, operational AI defence programmes and those still debating governance frameworks is widening rapidly.
Top story: The UK MoD awarded a £2bn, 15-year AI ‘Battle Lab’ contract to train 60,000 British soldiers annually — the largest AI-for-defence procurement commitment in UK history.
UK MoD Signs £2bn AI Battle Lab Deal to Train British Army
The Defense Post · Strategy
The UK Ministry of Defence has awarded a £2 billion, 15-year contract to Omnia Training — a Raytheon UK-led consortium — to build an AI-enabled ‘Combat Laboratory’ that will train up to 60,000 soldiers annually using AI, advanced analytics and virtual environments. The platform will replicate modern battlefield conditions, drawing explicitly on lessons from Ukraine, and will scale from formations of 100 to 50,000 troops. It is the most significant AI procurement commitment in UK defence history, signalling that AI adoption has moved from pilots to decade-long funded infrastructure.
https://thedefensepost.com/2026/07/13/uk-training-overhaul-omnia/
White House NSPM-11 Directive Puts AI Vendors on Notice Over Defence Contracts
Mintz · Regulation
The US White House’s National Security Presidential Memorandum 11 (NSPM-11) empowers agency heads to terminate contracts — for default or convenience — with any company that has ‘repeatedly demonstrated a pattern of conduct’ inconsistent with administration AI policy. The directive also bars the national security enterprise from using AI for unauthorised surveillance or embedding ideological bias into systems. Defence contractors, dual-use AI firms, and research institutions with national security affiliations must now urgently assess how these vendor accountability provisions affect their operations.
Australia Deploys German AI Surveillance Drone in Live Allied Exercise
Janes · Tools
The Australian Army has begun operational use of the Vector AI — a fixed-wing VTOL UAV from German firm Quantum Systems — describing it as its ‘newest’ surveillance platform. The drone was deployed during Exercise Southern Jackaroo 2026, a trilateral exercise with Japan and the United States, where it used AI-based software to map environments in real-time and detect and track targets to support artillery and UAV strikes. The deployment is a concrete example of NATO-adjacent allies fielding commercially sourced European AI hardware in live joint operations.
NATO Endorses Innovation Scale-Up Package, Issues First Industry AI Demand Signal
NATO · Strategy
At the 2026 Ankara Summit, Allied Leaders endorsed the NATO Innovation Scale-Up Package, which included the alliance’s first-ever public demand signal to industry and innovation ecosystems — a formal call to action directing private sector AI and technology firms toward specific capability gaps. The move reflects NATO’s shift from aspirational AI strategy to active procurement posture, with AI now embedded across cyber defence, intelligence analysis, drones and battlefield decision support. For European and UK AI startups, the demand signal opens a structured route into alliance procurement for the first time.
https://www.nato.int/en/what-we-do/deterrence-and-defence/emerging-and-disruptive-technologies
EUObserver: NATO’s Vital AI Infrastructure Is Built for Business, Not War
EUobserver · Risk
A sharp analysis published ahead of the Ankara Summit warned that NATO’s most critical AI infrastructure — cloud platforms, frontier models, data centres — was architected for commercial resilience, not wartime operational demands. The piece highlighted that in June 2026, the US government ordered Anthropic to restrict foreign access to its most capable AI models overnight for national security reasons, sending an unmistakable signal that Washington will treat frontier AI as a sovereign asset. This raises acute questions for NATO allies, including the UK, about dependence on US commercial AI for core cyberdefence capabilities.