A bridge that keeps stopping
South of Grantham, a 293-metre, 2,500-tonne bridge is inching across the East Coast Main Line and the River Witham at a few millimetres per minute. It is a remarkable piece of engineering — and it is running roughly five years late.
The Grantham Southern Relief Road was expected to be open to drivers by late 2023. The current estimate is 2028. Over that same period, the project's budget has roughly doubled, with costs now projected at £158–168 million. Behind those numbers lie not contractor disputes or planning battles but two distinct engineering errors, both centred on the same bridge.
The first — unstable ground on the eastern embankment, discovered in July 2022 — forced a redesign and added around 18 months to the programme. The second, revealed in February 2025, was a design flaw that made the intended method of pushing the bridge into place unsafe in certain wind conditions. Work stopped again.
The GSRR is not a uniquely cursed project. That is precisely what makes it instructive. What do two separate failures, on the same structure, within the same design and oversight process, reveal about how public infrastructure actually gets built — and who is held to account when it goes wrong?
The ground shifted: what the first delay exposed
In July 2022, engineers working on the eastern embankment noticed something wrong: the ground had moved. The soil beneath the planned bridge approach was too soft to carry the structure above it — a problem that, in an ideal process, would never have reached the construction stage at all.
Ground investigation is the stage at which engineers sample the soil, assess load-bearing capacity, and feed that data into foundation design. If the survey is incomplete, or if its findings are not adequately carried through into structural decisions, the gap tends to appear only when the ground behaves differently from what the drawings assumed. That appears to be what happened here.
Lincolnshire County Council left no ambiguity about where the responsibility lay. In a public statement, the council said: 'The coordination of the ground investigation information and the subsequent design of the bridge was and is the responsibility of the bridge designer, WSP.' Foundation design was WSP's remit; the council's position was that the failure to account for the unstable ground fell squarely within it.
The fix was not a simple stabilisation job. Engineers chose to extend the bridge itself, routing it past the problematic soil rather than attempting to build through it. That solution required a fresh planning application and added approximately £15m to a project then costed at £133m, while pushing the expected completion back by around 18 months — from late 2023 to 2025.
What the episode exposed was a coordination gap: between what site survey data recorded and what the foundation design ultimately assumed. Such gaps are a known risk in large civil engineering schemes. What made this one unusually visible was the combination of public money and a council prepared to name the designer directly.
The wind problem: a second failure and a broken assurance chain
Three years after the ground investigation failure, a second problem stopped the project — and in some ways it is more troubling than the first.
By February 2025 the bridge was structurally complete and ready for its launch: the incremental push method by which engineers slide a pre-built deck horizontally into position over the obstacle it is designed to cross. It is a technique used precisely to avoid lifting heavy structures across live railways. But before the push could begin, engineers established that the planned method was unsafe under specific wind conditions. Work halted while designers developed a revised approach.
The timing mattered. The project stood at £148m; resolving the wind problem added a further £10–20m, lifting the expected total to £158–168m — roughly double the original budget.
What amplifies the damage is that this error was supposed to be uncatchable. Following the first delay, all design work had been subjected to a mandatory, additional layer of independent checking. Lincolnshire County Council acknowledged plainly that 'despite all of this, a mistake was made.' If independent review did not catch the flaw, it raises a harder question than whether the original design was wrong: it asks whether the checking process was genuinely rigorous or had become procedural.
Main contractor Galliford Try moved quickly to clarify its position, stating it operates on a 'construct-only basis' and that responsibility for delays lies with bridge design. No designer has been publicly named for the wind-conditions error; the council has framed it as 'supplier error' and confirmed that contractual and legal processes are under way.
Accountability when public money is on the line
Public infrastructure occupies a different accountability register from private construction, and the GSRR makes that visible in concrete terms.
The project draws on a mix of national transport grants — including £11.9m from the Local Transport Board and £16.1m from the Single Local Growth Fund — alongside Highways England contributions, developer money, and Lincolnshire County Council's own budget. That layered funding structure means every cost overrun carries political weight: local taxpayers and the councillors representing them have a legitimate stake in understanding where the money went and who is liable.
Both delays produced immediate accountability responses. The council committed to covering extra costs in the first instance but pledged to pursue contractual and legal processes to recover 'every penny of taxpayer money' from those responsible. South Kesteven District Council's Finance and Economic Overview and Scrutiny Committee reviewed progress in February 2026, reflecting the multi-tier governance that a project of this scale attracts.
The council's willingness in 2022 to name WSP directly — in a public statement, during a live commercial relationship — is notable. Organisations managing large construction disputes routinely prefer careful language that does not prejudice legal proceedings. Naming a designer publicly is a departure from that norm, and it reflects the particular pressure that public money creates.
That same pressure has limits. Ongoing contractual and legal processes constrain how much detail can be disclosed about the second failure. The gap between what the public funding relationship demands — transparency, explanation, visible consequence — and what active litigation permits is a structural tension that no amount of political will can fully resolve.
How the bridge is crossing the valley now
Since March 2026, the bridge has been moving. The revised launch is proceeding at only a few millimetres per minute — a pace that is not a failure of ambition but a deliberate condition of the redesigned method. Every push over the railway happens during overnight Network Rail possessions, windows in which train services have stopped and the electrified lines have been disconnected. When morning approaches, work pauses and the tracks reopen.
While the structure moves, technical instruments monitor several points along its length simultaneously, every second, detecting any misalignment before it compounds. The monitoring intensity is itself a legacy of what came before: two failures have made precision non-negotiable.
As of late June 2026, the 293-metre structure was approximately one-third of the way across the valley. Engineers expect it to be fully in place by autumn 2026. The road itself — the reason the bridge exists — is scheduled to open to drivers in 2028.
The project is not finished, and the timeline carries the weight of everything that delayed it. But the bridge is, measurably, crossing.
What this project tells us about building in public
Two engineering failures, one bridge, and a quality-assurance system designed to prevent exactly that kind of repetition: the GSRR offers a concentrated version of a tension running through all large public infrastructure. Projects of this complexity carry risks that formal processes reduce but cannot eliminate. Ground conditions shift, structural dynamics produce surprises under specific wind loadings, and no pre-construction survey models every contingency.
The second failure sharpens that picture. When independent checking misses a design flaw, the question is not only whether the design was wrong — it is whether the checking was genuinely sceptical. A review that works through the same assumptions as the original design will miss the same things. The mandatory additional checking layer apparently did not provide that independence, and that finding stands regardless of what legal proceedings subsequently determine about liability.
Public accountability operates on a different timescale from engineering reality. Scrutiny committees, cost-recovery pledges, and the public naming of designers serve a legitimate democratic function, but they arrive after errors, not before them. The gap between the transparency that public money demands and the silence that active litigation imposes is real — no political will fully resolves it.
When the road opens in 2028, lorries currently threading through Grantham's town centre will have an alternative route, and the A1 junction south of the town will be more directly accessible to local drivers. That purpose is delayed but unchanged. The more durable lesson is narrower: independent checking in complex infrastructure works only when it is designed to find failures — not to confirm that the original designers found none.