HS2: How a £37.5bn National Railway Became a £100bn Stub
HS2 is the clearest live example of the Iron Law of Megaprojects in Britain. It was sold in 2009 as a £37.5bn national network reaching Manchester and Leeds. The northern legs were cancelled, yet the surviving London-Birmingham section alone is now estimated at £87.7bn to £102.7bn, and first trains have slipped from 2026 to somewhere between 2036 and 2039.
How much has HS2 gone over budget?
The full HS2 network was estimated at £37.5bn (2009 prices). After the Manchester and Leeds legs were cancelled, only London to Birmingham remains, and in May 2026 the Department for Transport estimated that single section at £87.7bn to £102.7bn (2025 mixed price base). That is roughly double its own 2020 funding envelope of £44.6bn, and up to about 2.7 times the original full-network figure for a fraction of the route.
The cost history, step by step
HS2's numbers are easy to misread because the price base and the scope both keep changing. The honest way to read them is to hold the two comparisons separate: what the whole network was meant to cost, and what the surviving Phase 1 was meant to cost on its own.
| Date | Scope | Estimate (price base) |
|---|---|---|
| 2009 | Full Y-network (London-Birmingham-Manchester-Leeds) | £37.5bn (2009) |
| 2013 | Full network | £50.1bn (2011) |
| 2015 | Full network (funding envelope) | £55.7bn (2015) |
| 2019 | Full network (Oakervee / Cook stocktake) | £72bn-£98bn (2015) |
| Apr 2020 | Phase 1 (London-Birmingham) Notice to Proceed | £40.3bn target, £44.6bn envelope (2019) |
| Oct 2023 | Phase 2 (Birmingham-Manchester) cancelled; London-Birmingham only | - |
| May 2026 | Phase 1 (London-Birmingham) only, post-reset | £87.7bn-£102.7bn (2025) |
Two things stand out. The route got shorter, losing the entire northern half, while the cost of what is left more than doubled its 2020 funding envelope. And the single London-Birmingham line now costs more than the whole national network was ever supposed to.
Why the cost rose
The National Audit Office and successive government reviews point to a consistent set of causes, most of them textbook optimism bias and strategic misrepresentation of the kind reference class forecasting is meant to correct:
- Politically low initial estimates. The early figures were set to secure approval, not to survive contact with the engineering. This is the same pattern the NAO identified on Crossrail and the Big Dig.
- Underestimated works. Tunnelling, viaducts such as the Colne Valley (Britain's longest railway bridge), and the scale of environmental and planning mitigation all cost far more than the business case assumed.
- Delivery inefficiency. The May 2026 reset attributed a large share of the increase to inefficiency within HS2 Ltd and its supply chain, alongside weak cost control.
- Inflation not captured. Earlier forecasts did not consistently reflect construction inflation, so real cost growth was masked until it was re-based upward.
A fundamental reset of the programme began in June 2025 under a new chief executive; the May 2026 Accounting Officer assessment is the first full re-baselining to come out of it.
What HS2 teaches about budget overruns
HS2 is a live illustration of Bent Flyvbjerg's Iron Law of Megaprojects: over budget, over time, over and over again. It also shows why the choice of baseline matters so much. Cutting scope did not cut cost, because the fixed early works, the tunnels and the reset overheads all stayed. Anyone forecasting a project of this class should have started from the reference class of comparable rail megaprojects, where a 40%-plus real cost overrun is the base rate, not the worst case.
Sources
- DfT (May 2026). HS2 Phase 1 Accounting Officer Assessment (£87.7bn-£102.7bn; opening 2036-2039).
- UK NAO (2024). HS2: update following cancellation of Phase 2.
- House of Commons Library (CBP-8601). HS2: the business case, costs and spending.
- Institute for Government. HS2: costs and controversies (cost timeline).