Record settlement, real cut.

STFC has just told everyone, in detail, what the next four years of cuts actually look like — not a warning that cuts are coming, but the actual numbers: which facilities, which programmes, how much, and by when.

A boundary before I go any further. I work for UKRI, so this isn't a post where I comment on internal decisions, budget deliberations, or anything I only know because I happen to work inside the organisation. Everything below is drawn from what's already public — Prospect's own statements, UKRI's and STFC's own published material, and reporting anyone can read. It's not privileged, and it isn't UKRI's institutional view. I'm writing as a lay rep, not a spokesperson for my employer.

With that said: the headline figures and the lived reality point in different directions at the same time, and both are true, which is most of why this is hard to talk about honestly. The 2025 Spending Review produced what's been widely described as a record settlement for UK research and development — £86 billion across the UK out to 2030, £38.6 billion of that through UKRI, rising toward £10 billion a year by the end of the period. Read only that, and funding for UK science looks like it's climbing.

Read UKRI's actual cash allocation for 2025 to 2026 and it looks different: £8.8 billion, the same nominal number as the year before. Flat cash is a real-terms cut once inflation's in the picture, and UKRI's own share of total UK research and development spending has fallen anyway, from 54% in 2018-19 to 43% now — more of the national total is going elsewhere, and the research councils aren't seeing proportionately more of it.

STFC has it worse than flat. A lot of its costs don't move with ordinary inflation at all. International subscriptions — to CERN, to the European Southern Observatory, the memberships that keep UK physicists inside the big international facilities — are being fully maintained, and are still on course to rise around 19% in cost. Energy for facilities that never switch off doesn't get cheaper because a spending review was generous elsewhere. Set costs like that against a flat-to-falling budget and the result is arithmetic, not mismanagement: STFC's total cost base has to fall from just over £1 billion to around £877 million by 2029-30, a 15% reduction, because nobody actually funded the gap between what it costs to keep the current estate running and what STFC's been given to run it.

What that means for particle physics, astronomy, and nuclear research is where the numbers get genuinely confusing, and it's worth being precise rather than picking whichever figure suits the argument. STFC's own headline is a 2.7% cut overall — the figure that includes protecting postdoctoral researchers within those grants at 2025-26 levels, inflation included, and continuing PhD studentships. That's true, and it's also not the number individual projects were actually asked to plan around. STFC's executive chair, Michele Dougherty, wrote to the community earlier this year setting out an expected reduction to discovery research funding closer to 30%, with individual projects asked to model the impact of cuts at 20%, 40%, and 60%, including the point at which their work stops being viable at all. Both figures are real. They're just answering different questions, and it matters which one you're being told.

The same gap between headline and reality shows up at the facility level, across STFC's sites at Harwell and Daresbury. ISIS will run fewer experiments. STFC is looking for outside funding for the muon beamlines at ISIS and for the Lasers for Science Facility, and will close them if it doesn't find any. The Clara test accelerator at Daresbury is being mothballed. Boulby's operating budget is down 40%. Compute spending drops £10 million a year, accelerator and technology support £8 million, professional services £12 million. UKRI is putting in £135 million of transitional money over three years to soften the landing — a genuine and genuinely necessary intervention, and also not the same thing as solving the underlying problem, which is that STFC's fixed costs are still rising faster than its budget, transition fund or not.

That postdoc protection is real, and I don't want to wave it away — but it sits inside a much wider picture where hundreds of posts are going regardless, some through voluntary exit schemes and some through redundancy as programmes wind down. UKRI's chief executive, Ian Chapman, has pointed out, fairly, that STFC hired around 800 people over the previous five years, and that context matters. It doesn't change what it's like to be one of the people whose post isn't there any more, or one of the people left to keep a facility running with fewer colleagues than it had last year. Protecting a grant line and protecting the people who'd have staffed the work around it are not the same thing, and both are true at once.

Sue Ferns, Prospect's senior deputy general secretary, has made the underlying point better than I can: the UK's standing as a research partner rests on national capacity — infrastructure and the skilled workforce behind it — not on the grant total alone, and as she put it, "denuding any of those diminishes our standing on the global research stage."

None of this reads to me as STFC mismanaging its own money, and I don't think that's a fair account of what's happened. Dougherty has been open that the process has been a difficult one for the people going through it, and I think that's genuinely meant, not a line. The harder truth sits a level up, in how the settlement itself was built. A spending review can produce a number that sounds generous against the national total while still landing as a real cut on the one council whose costs happen to be the least inflation-shaped of any of them — international subscriptions, ageing physical infrastructure, energy-hungry facilities that can't simply run fewer hours without losing the science they exist for. Prospect's general secretary, Mike Clancy, has talked about UK science funding settling into a boom-and-bust pattern, and having watched a few of these cycles now as a rep, I think that's an accurate description rather than a rhetorical one: a lean stretch, a settlement that reads as recovery in the press release, and then the same structural gap reopens a year or two later because what actually drove it — costs that were never going to track ordinary inflation — was never addressed the first time either.

It sits on top of something longer-running, too. Public sector science and technology pay, across UKRI and comparable bodies, has fallen by around a quarter in real terms since 2010. That's not this year's story, but it's the floor everything else stands on, and it's a large part of why people leave rather than wait out the next cycle — not dramatically, not all at once, just steadily, the way water finds a lower level.

What Prospect's actually asking for isn't complicated, and it isn't a blank cheque either: an honest account of what a settlement really buys once inflation and rising fixed costs are netted off, a proper assessment of the impact on people before decisions land rather than after, and some recognition that skills lost this year don't come back quickly, or sometimes at all, once the people carrying them have gone elsewhere. That last part is the one that actually sits with me, more than any of the figures above. A research software engineer or an accelerator physicist who leaves for industry, or leaves the country, isn't a line that un-cuts itself the next time a spending review is kinder.

There's a wrinkle over all of this that nobody can plan around yet. The UK has a new Prime Minister this week, Andy Burnham, and in time, presumably, new ministers to go with him. Prospect's already called on him to urgently review the STFC decision, which strikes me as exactly the right thing to ask for. What I can't tell you is what a change like that actually does to any of this — whether a new PM and a reshuffled science department reopens a settlement like this one, or whether it's already too baked in to move. Nobody I've spoken to seems to know either, and I'd rather say that plainly than pretend otherwise.

I don't have a tidy way to close this out. I've stepped back from my actual job for the next few months to do this full time instead — to be a rep, not a research software engineer, until we're through the other side of it. I don't like that. Not one bit. It's not how I wanted to spend this year, and it's not a decision I'd wish on any colleague who ends up having to make it too. But it felt necessary, and I'd rather do this properly than do it at the edges of a day job while people trusting me to represent them get less of me than they need.

What being a rep looks like through a year like this is mostly the same quiet question, asked in a lot of different phrasings, by a lot of different people, none of which I can actually answer yet: am I going to be alright? I don't know yet either. I know which answer I'm hoping for.