The claim is speed. The evidence on moving the furniture is that it costs you a year.

Start with the honest version of the pitch, because there is one. A Labour source briefed that the technology brief needs to be mainstreamed, that the department "has not been firing on all cylinders," and that putting science and technology inside a powerful, growth-focused business department would give it more weight, not less. That is a real argument. The Institute for Government, which studies these reorganisations for a living and is sceptical of most of them, has said plainly that some are worth doing: it judged the 2023 decision to reunite business and trade sensible on its own terms. Machinery can be wrong, and fixing it can be right.

So the question is not whether reorganising Whitehall is ever justified. It is what this particular one would cost, set against what it is meant to buy.

The measured part is where the marketing goes quiet. The Institute for Government's own costing puts the up-front bill for a machinery-of-government change at roughly £15m to £50m, in new IT and HR systems, security passes, office space and rebrands, before anyone has done a day of new work. The larger cost is the one ministers rarely feel: a conservative estimate the Institute uses is that a fifth of staff lose a fifth of their productivity for around ten months while the new structure beds in. Applied to departments the size of the ones involved here, that is tens of millions of pounds of work not done, on top of the direct bill. Its verdict on the whole category is one line: machinery-of-government changes are "only rarely worth the cost, disruption and distraction."

Now set that against the thing being reorganised. The Department for Science, Innovation and Technology houses the AI Security Institute, the Sovereign AI Fund, UK Research and Innovation and the Government Digital Service, and controls a budget of roughly £13.9bn, most of it UKRI's research money. And set it against the clock the government has set itself. The outgoing science secretary, Liz Kendall, said before her exit that Britain had roughly two years to establish a foothold in AI or have the technology "done to us." The tech industry's own estimate, reported this weekend, is that unpicking the department and redistributing it would take about six months. A quarter of a self-imposed two-year window, spent on the reorganisation rather than the thing.

That is the gap, and it is worth sitting in without forcing it shut. The want is real: a government that feels behind, reaching for the lever that looks like decisive action. The evidence is also real, and it points the other way. On the record, in Britain, moving functions between departments has usually bought delay first and benefit later, if at all. New structures, as the Institute puts it, are no substitute for collaboration. Whether this reorganisation is the rare one that clears the bar is not yet knowable, because the plan is not yet signed. What is knowable is the price of admission, and it is not speed.

Why it matters here: This is UK money and UK capacity, and the bill for redrawing the org chart falls on the very departments meant to be racing a two-year clock. Before treating a reorganisation as momentum, it is worth asking what the last few Whitehall reorganisations actually delivered, and when.