We Said the MOD Couldn't Afford the Old Way. The Defence Investment Plan Proved It.

A fortnight ago we argued that the Ministry of Defence could no longer afford to build software the way it always has. We wrote it while the Defence Investment Plan was still unpublished, and we set out four things that had to change if the plan was to deliver capability rather than just commitments. The plan has now landed. This is the reckoning.

We will resist the political lens. Whether 298 billion pounds is enough, whether the timing is right, whether the settlement is a legacy or a liability, those arguments are being had loudly elsewhere. We are interested in the strategic question underneath. Can the United Kingdom turn this money into capability at the pace the threat demands? On that question, the Defence Investment Plan is revealing.

More Money, Same Gap

The headline is real. The plan commits 298 billion pounds over four years, an additional 15 billion of funding, and puts core defence spending on a path to 2.7 percent of GDP from 2027. By any measure it is the largest sustained uplift since the Cold War.

And yet the gap we described in June has not closed. The settlement is backloaded, with the heaviest funding arriving in the later years, while the pressure to be ready sits in the first two. That is not our characterisation. It is the reason a Defence Secretary resigned in the run-up to publication, warning that the plan fell short of what the moment requires and reached only 2.68 percent of GDP by 2030, when the stated ambition had been 3 percent. More money is going in than at any point in living memory, and the affordability gap is still open.

This is precisely the point we made a fortnight ago. You cannot spend your way across a gap this structural. The arithmetic does not work, even with a historic uplift. Which leaves one lever nobody wants to pull first: building capability for materially less, materially faster. Efficiency is no longer a virtue to aspire to. It is the plan's load-bearing assumption.

The Scorecard

In June we set out four changes that would unlock savings within the plan's own lifetime. Here is how the Defence Investment Plan measures up.

Ring-fence funding for AI-accelerated delivery. Done, in principle. The plan includes a 500 million pound Transformation Fund explicitly for productivity-improving investments in AI and for reducing the department's dependence on consultancies. That is close to a verbatim statement of the case we have been making for a year. The caveat is that the money is not yet ring-fenced for governed, SME-led delivery of real programmes. The intent is now policy. The mechanism still needs building.

Use and reform the procurement routes. Better than we asked. Rather than simply working the existing frameworks harder, the plan restructures buying around segmented commercial pathways, including a dedicated Digital and Technology route and, for the first time, a specific SME pathway designed to simplify contracting and cut administrative burden. The routes for agile, iterative delivery are being widened, not merely tolerated.

Demand governance, not prohibition. Partially. The cultural reflex that treats AI as a risk to be banned rather than a capability to be governed is still present across Defence. But the appetite is now funded, with 115 million pounds allocated to strengthen the country's defences against AI risks. The question we posed stands. The answer to AI security concerns is assurance, not abstinence. Governed adoption, with cleared human review at every stage, is the operational response, and it already exists.

Engage the UK SME sector. Reinforced. The plan restates the commitment to grow defence spending with SMEs by 50 percent to 7.5 billion pounds by 2028, and layers on a "buy British by default" approach underpinned by a clear definition of a British company. For a UK-owned, security-cleared SME building governed AI delivery today, this is the language of the plan written in your favour.

The Gap the Plan Doesn't Close

Three of our four asks are visible in the text. That is encouraging. It is not the same as delivery.

The plan's central weakness is a timing mismatch. The threat is front-loaded and the money is backloaded. Russia is testing British defences now. China is scaling AI-enabled military capability now. Neither is waiting for the later years of a UK spending profile. The bridge across that mismatch cannot be more money arriving sooner, because it is not. The bridge is delivering more capability per pound in the years when the pound is tightest.

That is what governed AI-accelerated engineering is for. Weeks instead of years. A fraction of the traditional cost. Sovereign ownership of the code, full audit trails, cleared engineers, and security assurance built in from the requirement rather than bolted on at the end.

There is a wrong way to chase this, and the funding pressure will make it tempting. Ungoverned acceleration, code generated at speed with no assurance layer, no review, no accountability, is not a shortcut. It is how you manufacture the next black-box incident and the next lessons-learned review. Speed is not the risk. Ungoverned speed is. Governed acceleration is the only version Defence can actually deploy.

The Strategic Reality

The Defence Investment Plan is, in the end, a set of choices. Every pound committed to one capability is a pound unavailable to another, and the plan is candid that it was costed line by line to expose exactly those trade-offs. The most consequential choice in it is not a platform or a programme. It is the decision, half-made and under-resourced, to use AI to build faster and for less.

The plan now says as much in its own funding lines.

We wrote in June that the old way of building defence software was a luxury the MOD could no longer afford. The Defence Investment Plan has just agreed. The open question is the same one we ended on then, only now it is urgent rather than rhetorical. Not whether governed AI-accelerated development becomes the norm in Defence, but whether the UK adopts it fast enough to matter.

---

Signal Coding is a UK-owned AI-accelerated software engineering company serving Government, Defence, and National Security. We deliver governed AI development with defence-grade security assurance: human-in-the-loop oversight, SC/DV cleared engineers, and full audit trails from requirement to deployment. Contact us to discuss how AI-accelerated delivery can close the gap between what your programme needs and what it can afford.