Construction AI: Gateway 2 lets go of the cables, and Ofgem charges rent on the grid queue

So here's a quiet one that slipped out under the noise of the autumn deadlines. From 1 September 2026, a slice of building work stops passing through Gateway 2 at all. The government confirmed it in Circular 02/2026, made using section 11 of the Building Act 1984, and it covers telecoms: the installation of fibre optic cabling, and building work to mobile communications masts on rooftops.

What that means in practice is this. For fibre optic cabling, the procedural requirement to get building control approval before you start work is dispensed with, and it applies both to buildings inside the higher-risk regime and those outside it. For rooftop masts, the relief runs at Gateway 2 only for buildings in the higher-risk scope, and Gateway 3 still bites. The fibre carve-out is time-limited to three years, so it's a trial with a clock on it rather than a permanent hole in the regime.

I'll say plainly I think this is the right sort of pruning. Gateway 2's backlog has been the sector's slow puncture all year, and pushing a fibre pull through the full higher-risk process was never a proportionate use of anyone's fortnight. But the comparison only goes so far. Less red tape helps the programme only if the record still holds, and a dispensation from the front-door check doesn't dispense you from the golden thread. Someone still has to log what went in, where, and to what standard, because the day a fire-safety query lands you'll want that on file. That's the bit to hold onto.

Practical bit: if you've telecoms work queued behind a Gateway 2 submission, check whether it now falls outside the door entirely from 1 September, and put the three-year fibre window in your programme diary so nobody's caught out when the check returns.

We flagged on 1 September that the wall in front of the UK's £100bn data centre pipeline had moved from the planners to the grid. On the same theme, and worth the update, the regulator's now acting on exactly that. Ofgem's consultation on a data centre commitment fee closes on 16 September 2026, and it's the clearest attempt yet to clear the connection queue of projects that are only holding a slot in case.

The mechanics are blunt in a good way. A large-load data centre would pay a fee when it accepts a connection offer, refunded once the project energises and forfeited if it exits the queue early. Ofgem's proposed range is £237,500 to £712,500 per megawatt, which it reckons is roughly 2.5 to 7.5 per cent of a scheme's cost. Those are the regulator's figures, so treat the percentages as an argument rather than a fact. Underneath sits the number that explains the whole move: connection applications have jumped from 41GW to 125GW in under a year, and at least 80GW of that is data centres. The commercial press has taken to calling it pay to stay.

So what does this do for a contractor? It thins the pipeline you're bidding into, and it should thin it towards the schemes with real money and a real connection date behind them. What I like about it is that it rewards the developer who's actually ready to build, which is the person you want to be working for anyway. A speculative land-bank with a grid slot was never going to keep a site team in work for long.

The procurement filter: before you commit bid resource to a data centre framework, ask whether the developer has paid to stay, because a fee they'd forfeit is a cheaper signal of intent than any amount of glossy prospectus.

Here's a number I keep turning over, from McKinsey's State of AI 2026 survey, reported in late August. Roughly a third of organisations, 32 per cent, said they'd decided against buying an off-the-shelf software product this year and built their own with agentic coding tools instead. Among the high performers, the firms attributing real profit to AI, it's nearly half. McKinsey also notes the actual EBIT impact stayed flat at 37 per cent, so nobody should read this as a productivity miracle landing tomorrow.

But the direction matters more than the ROI line, and it's the mechanism behind a category I keep pointing at: challenger construction software, the broad AI-first platforms taking on the big incumbents the way the challenger banks took on the high street. What's happened is the cost of building software has collapsed. A small team can now ship the breadth that used to take an incumbent's whole product suite and the best part of a decade. That's why the incumbents' pricing is starting to look exposed. Procore's own filings show net revenue retention of around 114 per cent, which is a polite way of saying existing customers paid about 14 per cent more year on year, and Autodesk lifted list prices 6 to 9 per cent in January and began trimming renewal discounts. Those figures are as reported, but any commercial director who's signed a renewal lately will recognise the shape of them.

I'm not going to pretend a mid-size contractor is about to code its own project platform over a wet weekend. The comparison only goes so far. But the same force that let a third of firms skip a purchase is the force letting founder-led tools cover a whole project, not one task, at a published price you can walk away from. That's what it's about. When your incumbent renewal lands 14 per cent higher this year, you've now got somewhere else to look.

For your board pack: put your two biggest software renewals on one slide with their year-on-year increase, and next to each one name a credible AI-first alternative. The exercise alone changes the next negotiation.

Three moves this week, one shape. The government's taking a procedural check off telecoms work, the regulator's charging rent on grid slots, and the market's quietly deciding it can build cheaper than it can buy. Each one shifts a cost or a gate. None of them changes the thing that actually protects you, which is a record you can stand behind when someone asks.

So the standing discipline holds. Cheaper software, lighter regulation and a thinner queue are all worth having, and they all land on the same person in the end, the PM or the site manager who has to show what was done and why. Get the information straight first. Everything clever sits on top of that, or it sits on sand.

A practical step: pick one live job and ask whether its document trail would survive a hard question at the next Gateway. If the answer's no, that's this week's task, before any new tool goes anywhere near it.