Is Britain's Military Winning The War In Cyberspace?
7 min read
The UK’s military cyber command recently marked its first anniversary. Can it meet the challenges of an increasingly complex battlefield? Noah Vickers reports.
Britain’s army, navy and air force have each come under the political spotlight over the last year amid rows over military funding which led to John Healey’s resignation as defence secretary and Sir Keir Starmer’s exit from Downing Street.
But less discussed in Westminster – and almost completely unknown to the public – is a fourth wing of Britain’s military, which has undergone significant change following Starmer and Healey’s 2025 Strategic Defence Review (SDR).
It is called the Cyber and Specialist Operations Command (CSOC). Previously known as Strategic Command, its new name was chosen to reflect the growing importance of cyber and electromagnetic (CyberEM) warfare on today’s battlefield.
While responsibility for CyberEM was previously fragmented across several different parts of the UK’s armed forces, the establishment of CSOC in September 2025 has cohered those operations into a single command centre.
Though still in its early days, CSOC’s Major Neil Donaghy tells The House the shake-up is already delivering positive results in the invisible war being continually waged across cyberspace.
“We’re seeing activity daily and competition is constant nowadays,” he says. “Our ability to detect, respond, recover, co-ordinate that approach, specifically in cyberspace, has been tangibly improved, because we have ownership and control of all those capabilities and force elements to do that.
“Whereas before we didn’t – it was pockets of ecosystems and federations in bubbles which didn’t necessarily speak or co-ordinate together.”
The SDR recognised CyberEM as the “enabling domain that integrates all others” in modern conflict. While warfare has for more than a century involved attempts to jam or spoof opponents’ use of the electromagnetic spectrum, the rise of digitally enabled weaponry has opened up new avenues of attack aimed at restricting an adversary’s decision-making capacity or degrading their firepower.
“Pretty much everything on the battlefield these days, apart from the humble rifle, is digitised in some shape or form, so you already see the potential of introducing malicious code into all these different systems and capabilities, and the effect that it can have,” says Dr Thomas Withington, an associate fellow at the Royal United Services Institute (RUSI).
He points to Russia’s 2014 invasion of Ukraine as one key example of a conflict in which cyber tactics appeared to prove decisive.
A Russian cyber espionage group called Fancy Bear – thought to be part of Moscow’s GRU military intelligence service – created a piece of malware known as X-Agent targeted at devices which use the Android operating system.
According to a report by US cybersecurity firm Crowdstrike – the findings of which were denied by Ukraine’s defence ministry – the Ukrainians were using a software application on Android-based devices to manage their artillery fire control on their eastern frontline.
Over several months, Fancy Bear tricked Ukrainian soldiers on military forums into downloading a version of the application infected with X-Agent. Those devices were then compromised, granting the Russians access to their text messages, call logs and contacts – all of which would have formed useful intelligence.
But while that caused damage enough, the malware also enabled the Russians to triangulate the rough location of each infected device.
“Once you find the device, you find the soldier,” says Withington. “They were able to do very effective counter-battery fire against the Ukrainians, working out where their artillery placements were and attacking them.
“It’s thought that X-Agent played a key role in the degradation of Ukrainian artillery during that initial invasion. That was an example where it [cyber warfare] had a fairly profound effect.”
Britain’s own ability to compete in CyberEM is not without its challenges. Research by the Parliamentary Office of Science and Technology (Post) found last year that our military’s prowess in electromagnetic warfare (EW) has been hampered by “slow and complex Ministry of Defence (MoD) procurement processes, a lack of skilled personnel and insufficient sovereign capabilities”.
MoD contracts for EW-related technologies “often focus on development and delivery over five-to-10-year periods, longer than the fast-paced innovation cycles often observed in wartime”, the research points out. There is also a lack of specific higher education pathways in the field, as the UK has just one university – Cranfield – which offers specialist EW postgraduate courses.
Concerns have also been flagged by the House of Lords International Relations and Defence Committee over the UK and Europe’s limited sovereign EW capabilities and consequent reliance on the US.
“Once you find the device, you find the soldier”
Post’s research notes: “Industry and academic stakeholders have raised concerns that this reliance may prevent access to the most advanced technologies, for example, if the US decided to reserve its most sophisticated capabilities for itself.”
Donaghy pushes back on this point, saying that the UK “does have significant capabilities” even if “we haven’t tested them in conflict as much as Russia” and that the Defence Investment Plan includes £2.5bn over the next four years on CyberEM alone, with a further £5bn allocated from 2030 through to 2035.
The MoD is reforming its procurement processes, he adds, while efforts are also being made to gain and sustain skilled recruits in CyberEM professions, including through a Cyber Direct Entry Scheme designed to nurture talent in the field.
Yet the biggest challenge may be the arrival of increasingly sophisticated AI models and their ability to seek out vulnerabilities in legacy systems. While much of the public and media is focused on threats to civilian and economic infrastructure, Britain’s military technology is subject to some of the same risks.
“The pace of the threat is exponentially growing,” says Donaghy. “The frontier models have changed the game. They have put capabilities into the hands of hacktivists and small users.
“With £3,000 or £4,000, a local language model and a bit of training, you can put an agent on task and let it go and search, hunt and attack networks.” A local language model is an AI agent that can operate independently of remote servers.
On a visit to Parliament, CSOC’s commander General Sir Rob Magowan reveals that the UK military is seeking access to the latest models developed by US firms Anthropic and OpenAI.
“Frontier AI models offer the opportunity for an adversary to identify weakness in networks at a speed and scale that’s never been seen before,” he tells The House.
“Therefore our ability to understand those models, to ensure that we can protect ourselves against them, is obviously critical.”
As to whether such models could be used offensively as well as defensively by the UK, he says: “Our ability to prosecute effect using them is something that we would wish to explore and to see what the opportunities are, but it’s just the beginning of the conversation.”
Referring specifically to Anthropic and OpenAI, Sir Rob adds: “They’re US companies, they are very carefully regulated in the United States, and therefore the key is for us to stay close to our US colleagues, to understand how we can gain access to these models, so we can understand and protect ourselves against them.”
China, he notes, has been “more prolific” in developing models which may not be strictly as powerful as their US counterparts, but which are released onto the open market, allowing cyber-attackers to utilise them.
“We think that’s worrying, because of their potentially compromising effect and therefore we need to have a conversation – an international conversation – about how we manage and control these powerful frontier models,” he says.
In October, Sir Rob also spoke at the Resilience Conference in London alongside GCHQ director Anne Keast-Butler, where the latter warned that frontier AI is “starting to really expose vulnerabilities in legacy tech that we’re all sitting on”.
Asked whether he was conscious of a similar risk to the systems used by Britain’s military, Sir Rob replies: “Of course. Anne actually made the point on stage after I’d spoken about frontier AI that we’ve absolutely got a challenge against these powerful models, but that doesn’t excuse the fact that we’ve got to make sure we shut the front door and the back door. Basic security of our networks today, basic hygiene processes, remain important.
“So it’s both – it’s fixing the basics, it’s ensuring the foundations are as secure as they can be, but also understanding some of these powerful models.
“Are our networks wholly secure today? No, they’re not. But we are really investing in our architecture to ensure we can get much better at securing these networks.”
The MoD’s Digital Backbone project, he adds, will assist in “identifying where those vulnerabilities are in our legacy networks and getting after them”.