The past week has seen a convergence of AI becoming operationally embedded across Western defence at speed — from the UK’s MoD deploying AI targeting systems on the battlefield, to the Pentagon funding swarm-orchestration software that keeps drones flying through jamming. Meanwhile, the European Parliament is sounding the alarm that the AI Act leaves military systems ungoverned, creating a regulatory blind spot just as EU defence spending exceeds €400bn.
Top story: UK’s MoD Dstl reveals Projects VANAHEIM and ANVIL — AI computer vision targeting and counter-drone systems designed to move from experimentation into deployable British Army capability.
UK MoD Deploys AI Targeting and Counter-Drone Projects for British Army
Army Recognition · Strategy
The UK Ministry of Defence published a Dstl case study on 27 July outlining two live programmes: Project ANVIL applies AI and computer vision to help British Army drones locate and identify targets, while Project VANAHEIM focuses on protecting troops from hostile drones. Both are designed to translate battlefield lessons from Ukraine into deployable capability, signalling a sharp acceleration in the UK’s willingness to field AI-enabled lethal systems.
Pentagon Awards $10M to NODA AI for Jam-Proof Drone Swarm Software
Defence Blog · Generative AI
NODA AI secured a $10 million US Department of War contract on 22 July to build MAESTRO — software that translates a commander’s intent into coordinated autonomous action across air, land, sea, and subsurface drones from any manufacturer, even when communications are severed by jamming or cyberattack. The vendor-agnostic design is a deliberate move by the Pentagon to avoid hardware lock-in, and the contract integrates directly with LUCAS, the low-cost attack drone reverse-engineered from Iran’s Shahed-136. For defence practitioners, it marks a tangible step toward AI-orchestrated multi-domain autonomous warfare.
https://defence-blog.com/noda-ai-wins-10m-deal-to-make-pentagon-drones-work-as-one-team/
British Army to Train 60,000 Soldiers a Year via £2bn AI Combat Lab
National Security News · Tools
A 15-year, £2 billion contract has been awarded to Omnia Training — a Raytheon UK-led consortium — to build a digital Combat Laboratory that uses AI, advanced data analytics, and virtual battlefields to train up to 60,000 British soldiers annually. The platform can simulate operations from 100-person units up to 50,000-strong formations, with AI analysing individual and unit performance before deployment. Crucially, the intellectual property is being kept in British ownership to reduce dependence on overseas suppliers.
European Parliament Warns AI Act Leaves Military Systems Ungoverned
EU Perspectives · Regulation
A joint workshop of the European Parliament’s Human Rights Subcommittee and its Security and Defence Committee (SEDE) raised urgent concerns that the EU AI Act, despite its landmark status, deliberately excluded military and national security AI from its scope — creating a governance vacuum precisely as EU defence spending surpasses €400bn and AI-enabled weapons proliferate. SEDE chair Marie-Agnes Strack-Zimmermann warned that Europe risks ceding the normative rules of AI-driven warfare to Washington, Beijing, and Moscow. The session signals growing legislative pressure to close the military AI exemption.
https://euperspectives.eu/2026/07/ai-at-war-eu-ai-act-military-exemption/
Russian State Hackers Targeting Western Governments via Collaboration Software
Enterprise Times · Risk
The FBI’s Internet Crime Complaint Center issued an alert in the week of 20 July warning that a Russian state-backed cyber group, designated LAUNDRY BEAR, has been compromising Western government and commercial organisations via the Zimbra Collaboration Suite since at least July 2025. The alert is a direct reminder to defence and security practitioners that AI-era threats still include persistent, state-sponsored intrusion campaigns exploiting widely-used enterprise software. Organisations are urged to audit Zimbra deployments immediately.