General Sir Roly Walker has put a hard number on what AI is already doing inside British Army planning: a corps cycle that once took 72 hours can now take one. You should read that as a procurement signal, not a laboratory boast.
Speaking at the RUSI Land Warfare Conference in London on June 23, Walker said the Army's ASGARD system has cut corps-level planning from three days to an hour. Business Insider reported the same speech and the number is the one that matters, because headquarters work is usually where speed goes to die. If a corps can move from finding targets to coordinating strikes in one hour, every allied army still buying command software on old timelines has a problem.
Walker didn't present ASGARD as a clean demo in a conference hall. He said the system can let a corps attack ten times as many targets in a day, with the limit coming from available munitions rather than staff work. He also described it as a "digital juggernaut" evolving every 8 to 12 weeks. That is the line you can imagine every defense-tech founder underlining, because it turns AI in command and control from a pitch into a named operational claim by the chief of the British Army.
The basic job of ASGARD is simple to state and hard to build. It gathers battlefield data, processes it, helps commanders identify targets, and coordinates attacks faster than a traditional headquarters can do by hand. Business Insider called it a "digital targeting web" and reported that the UK recently ran it from a tube station under Trafalgar Square while managing troops in Estonia. The temporary headquarters was processing about 10 terabytes of data a day, which the British Army compared with nearly three months of nonstop high-definition Netflix.
That detail matters because it strips away the romance. This isn't AI as a slide on future warfare. It's a staff under London, directing an exercise in Estonia, with enough data moving through the system each day to bury a human planning cell in sorting work before lunch.
The UK has been moving this way for a while. The government announced last year that it planned to spend £1 billion, about $1.3 billion, on AI-enabled military systems. The Times reported from the same RUSI conference that Walker wants the future force to follow a 20/40/40 mix: 20% traditional platforms, 40% expendable systems, and 40% reusable AI-driven assets. Tanks and artillery don't disappear in that model, but they stop being the whole story. Drones, sensors, software, and electronic warfare carry more of the load.
Walker also said the Army is sending thousands of drones to units and introducing 50 operational-level electronic warfare systems used in Ukraine. That Ukraine comparison is not decoration. It is the pressure behind the whole argument. Russia's war has shown what cheap drones, fast targeting loops, and constant adaptation do to armies built around slower processes. Any headquarters that still needs days to make decisions is giving the other side time it doesn't deserve.
For companies selling AI command systems, ASGARD is a gift, even where the British Army hasn't publicly laid out every vendor and integration detail. Anduril is the obvious American name investors will reach for because its Lattice software is built around connecting sensors, operators, and weapons. Helsing is the obvious European one because it started with battlefield AI software and has since pushed into drones and other autonomous systems. But the important point is bigger than either company. Once a NATO army says out loud that AI has compressed a 72-hour process into one hour, the buyer's question changes.
Frankly, the question is no longer whether military AI can produce measurable staff-level gains. Walker just answered that in public. The question is who can deliver similar systems quickly enough, safely enough, and with enough trust from commanders who still carry legal and moral responsibility for the strike.
Investors had already been moving before Walker took the stage. Business Insider reported in May that Anduril raised $5 billion at a $61 billion valuation, roughly double its value from June 2025. The Financial Times reported last month that Munich-based Helsing was set to raise $1.2 billion at an $18 billion valuation, up from the €12 billion valuation it reached after a €600 million round led by Daniel Ek's Prima Materia less than a year earlier. Those are not normal software multiples. They are a bet that defense ministries are going to buy fast, buy repeatedly, and buy from companies that can update systems in months rather than procurement cycles measured in years.
The contracts are already following. Investor's Business Daily reported in March that the US Army awarded Anduril an enterprise contract worth up to $20 billion over ten years, consolidating more than 120 existing procurement actions. The Financial Times reported in February that Helsing won an initial €269 million Bundeswehr deal for HX-2 loitering munitions, with options that could take the contract to €1.46 billion. Those numbers give you the shape of the market: software, drones, targeting, counter-drone systems, and the command layer that ties them together.
There is a hard caution here. A faster headquarters is not automatically a better headquarters. If the data is poor, if the model misreads a target, or if commanders become rubber stamps for machine-shaped recommendations, speed becomes a risk rather than an advantage. The Army's job is not to worship the one-hour cycle. It is to prove that the hour is reliable.
Still, Walker's RUSI speech will travel. Allied militaries now have a public benchmark from a named general at a named conference, tied to a system that has been exercised across borders. That is exactly the sort of proof defense buyers use when they start asking why their own planning tools still move at the pace of paperwork.
Also read: A San Jose legal-tech startup just sued the US government over losing Anthropic's Fable 5 and it won't be the last • SK Hynix is betting $29 billion that the AI memory boom is nowhere near over • ASE Technology raises its 2026 capex to $8.5 billion as AI chip packaging demand overwhelms supply
