For most of my professional life, building major weapons systems was a long business. Years of research and development, followed by trials and then, for many of the most significant capabilities of the late twentieth century, something close to silence. Whole classes of weapon were designed, tested, fielded and retired without ever being used in anger. That did not make a system worthless; it remained ready to be used. But the gap between the trials area and the battlefield was, for many programmes, never crossed at all.
If the last four years have shown us anything, it is that those days are gone. The earlier pieces in this series described one part of the change: development has moved from a shipbuilding cadence to a software one, with systems revised and improved continuously through life. The other part is starker. Weapons are now being designed in the expectation that they will be used in the near future or, in some theatres, almost immediately. Ukraine has compressed the distance between research, development, test and employment in combat from years to weeks, and sometimes less. A capability conceived in the spring can be shaping the fight by the summer.
A great deal is said at the moment about production at the pace of relevance. General Sir Jim Hockenhull has called the pace of relevance “an equation about how fast our adversaries are moving”: being substantially better in five years is immaterial if the advantage is needed now, or if the adversary has already adapted. I agree with that; the clock is set by the threat, not by the industrial timetable, and everything in the programme is being compressed to match it, including the time available to understand what the system will do when it is used. When the distance between the drawing board and the battlefield shrinks from years to weeks, the opportunity to find problems before they become effects shrinks along with it. And when effects go wrong in armed conflict, people who should be protected are sometimes harmed. That is first and always a human cost. It is also, for the company whose name is on the system, the beginning of a set of questions that can lead to severe consequences.
None of this reflects any lack of seriousness on the part of those who build defence capability. In my experience the opposite is true. The division of labour that governs this sector is long established and entirely reasonable: the State sets its requirement, the manufacturer builds to that requirement and the State reviews it. The legal obligation to review new weapons, means and methods of warfare absolutely sits with States, and decisions about specific employment of such systems sit with the operational commander. A manufacturer that relies on that division is not being careless. It is relying on an architecture that served the sector well for decades. My point is that the architecture was built for slower times, and manufacturers are potentially exposed to more risk than was previously true.
The clearest recent illustration comes from outside defence altogether. The Horizon system at the centre of the Post Office scandal was built and maintained by a supplier; the prosecutions that wrecked hundreds of lives were the operator’s. On the traditional division of labour, the storm belonged to the Post Office.
It has not stayed there.
The inquiry’s first report found that people in both organisations knew, or ought to have known, that the system was flawed while courts were being told it was robust. The supplier, meanwhile, has accepted a moral obligation to contribute to redress, has stepped back from bidding for new government work until the inquiry is complete, and it has seen former personnel drawn into the police investigation alongside the Post Office’s own. None of that required a court to find anything against it. That is the point worth noting: by the time formal findings arrive, the commercial and reputational consequences have long since been felt, and they fell on the supplier as much as the operator.
Now translate that into the defence sector, where the product is not accounting software but a weapon, and where the pace I described above means more use, in more places, with less time between design and employment. The exposures begin to stack up:
The first is reputational, and in this sector reputation has more constituencies than most. The public, whose view can harden within a news cycle. Your own people, who chose to work in defence in the belief that it is done properly, and the engineers you want to recruit, who have options. Investors, for whom association with an unlawful effect is a risk they did not price in. Campaign groups, who are organised, legally literate and patient: they attend AGMs, they challenge licensing decisions in the courts, they investigate supply chains, and a company that cannot show its working makes their case for it. And buyers, because a procuring State reads the same newspapers.
The second is regulatory, and the news is littered with recent examples. Last month Airbus Operations paid HMRC a compound settlement of £6.4 million, the largest ever reached for strategic export offences, after self-reporting that over a sustained period it had failed to keep the records its export licences required. The breach itself is worth noting. It was not shipping to the wrong destination; it was failing to keep proper records. HMRC concluded settlements worth nearly £3.7 million with three exporters in a single quarter of 2025, and sanctions breaches have drawn seven figure settlements, including against an exporter that never shipped to the sanctioned country at all; the breach lay in who the recipient was connected to. Those are the cases resolved without prosecution.
Then comes the statutory inquiry, and the Post Office years are the reason it belongs on this ladder in its own right. When a capability is connected to serious harm, the United Kingdom’s institutional reflex is an inquiry, which carries a unique kind of jeopardy. It can compel documents and witnesses. It works in public, over years, with a chair and counsel whose job it is to reconstruct exactly what was known, by whom, and when. Companies and individuals criticised in an inquiry report carry that criticism permanently, and the answer to a seemingly simple question shapes everything that follows: what did you know, and what did you do about it?
At the top of the ladder sits criminal law. It is settled that individuals in a supply chain can be convicted of complicity in war crimes; a Dutch businessman served a long sentence for supplying precursor chemicals for mustard gas to Iraq, a conviction upheld by the Dutch Supreme Court. Courts in Europe have shown themselves increasingly willing to examine companies and executives whose commercial decisions touched armed conflict, and English law has the machinery ready: war crimes and crimes against humanity are domestic offences here, ancillary liability reaches those who aid or abet, and a company is a person capable of committing both. That top rung is rarely reached, but the question asked at every rung is the same one.
There is a better position to be in, and it is the reason I wrote this piece. The answer to that question has to be built, and it is best built early, and everything the first three articles of this series described is the building of it. A company that engaged with the lawfulness of its product from the design table, put the legal questions into its trials, kept the dossier as the product evolved, and can show precisely what it knew and what it did about it, has an answer that works at every rung of the ladder.
It works with the journalist, the campaigner, the investor, the regulator, the inquiry and, in the worst case, the court. The same record that gets a capability adopted at the pace of relevance is the record that protects the company that built it. Those are not two pieces of work; they are one.
The work of building that record can be modest, and it is at its cheapest right now, while your programmes are moving at the new speed and the record can be built as you go. Set against the cost of any single rung of that ladder, it barely registers. If you would like to know what that record should look like for your programmes, my door is open.
This is the fourth in a short series for defence manufacturers, technology firms and those who bring military capability to market: what the law asks of a system, and how answering those questions early strips risk, cost and delay out of a programme. The first three made the case for legal input at the design table, for treating the review as a continuing relationship, and for evidence that is designed rather than hoped for. Next: the final piece in this series, on why I do this work, and what I can do for yours.
