Ground Data, Grid Gaps, and 200,000 Jobs

MP Mike Reader presented the Ground Data for Growth Bill to the House of Commons last week, and it passed its first reading. If it becomes law, it would mandate that all geotechnical data gathered from site investigations be contributed to a centralised national repository - with standardised formats and shared access for key stakeholders.

The current situation is the opposite of that. Subsurface ground data is gathered project by project, held in different formats by different firms, and rarely re-used. Every new site investigation largely starts from scratch, even in areas where dozens of investigations have already taken place. It's not just inefficient - it actively slows planning, increases risk assessments, and adds costs that flow through to housebuilding timelines and infrastructure delivery.

The backing behind this bill is notable. Arcadis, AtkinsRealis, Arup, Mott MacDonald, and the British Geological Survey are all publicly supporting it. These aren't peripheral voices in UK construction and infrastructure - they're the firms doing the work.

Why this matters for AI in construction: A centralised, standardised geotechnical data repository is one of the foundational ingredients for applying machine learning to planning approvals, ground risk assessment, and materials optimisation. You can't train models on data that doesn't exist in a structured, accessible form. This bill, if it passes, creates that foundation. It's the kind of upstream infrastructure reform that unlocks downstream AI capability - and it's rarely this visible at the parliamentary level.

The bill is at first reading stage. But it's worth tracking. The construction industry has a direct interest in seeing it progress.

DP World has launched a £36M procurement for Boxbay at London Gateway Port, the UK's only deep-water port on the Thames. Boxbay is a high-bay automated storage system for empty containers - the kind of vertical, AI-adjacent logistics infrastructure that delivers roughly three times the throughput of traditional crane-based systems.

The project scope is substantial: more than 5,000 precast piles, over 15,000 tonnes of structural steel, and 3km of rail for automated storage-retrieval machines (SRMs). Construction is scheduled from mid-July 2026 to mid-July 2028. Bids close on 29 April 2026.

For construction firms with civil engineering and structural steel capability, this is a live procurement. For the wider industry, it's a proof-of-concept for high-density automated logistics at UK port scale - the kind of built environment that will become more common as supply chain automation accelerates.

Bloomberg and Sightline Climate data published last week put a stark number on the US AI data centre construction market: nearly 50% of planned projects are being delayed or cancelled.

The cause isn't a shortage of AI chips. It's a shortage of electrical infrastructure components - primarily transformers and batteries - that are manufactured predominantly in China. US data centre capacity currently sits at around 4GW against an industry target of 12GW. Wait times for electricity transformers have stretched from two to three years in 2020 to five years as of 2026. Trump-era tariffs have compounded cost pressures across the board.

This matters for UK and European construction firms for a specific reason. The US supply chain crisis doesn't make global AI infrastructure demand disappear - it redirects it. European data centre projects face the same electrical component constraints, but the pipeline is smaller, the political environment is different, and pent-up demand from stalled US programmes will create pressure to accelerate European builds. Firms with existing data centre construction capability and supply chain relationships are well placed.

Goldman Sachs context: The same week, Goldman Sachs released forecasts showing that AI-related data centre construction has created more than 200,000 additional jobs in the US since 2022, offsetting AI-related layoffs elsewhere. Total US economic injection from AI infrastructure investment since 2022 stands at $325 billion. The construction demand side of AI infrastructure is not theoretical - it's measurable and growing.

The Information Technology and Innovation Foundation published a new report today (6 April) addressing five key concerns about AI data centres: electricity demand, grid capacity crowding, rising household bills, grid reliability, and water strain.

The core finding: the problem isn't AI infrastructure per se, but the frameworks used to manage its impact. ITIF's recommendation is grid framework reform and improved demand management, not capacity caps or moratoria on AI data centre development.

The report is relevant for UK planning. Ireland's EirGrid has warned of a potential "mass exodus" of data centres due to grid connection stalls, and Amazon has already paused Irish investment. The same grid connection pressure is building in the UK. What happens in Ireland, Germany, and the Netherlands will shape what UK planners and policymakers reach for. Getting ahead of that debate - particularly if your firm is working on AI infrastructure delivery - means understanding the policy direction before it becomes planning policy.

Construction News reported last October that a significant tech skills gap is hindering AI uptake in UK construction. Four in ten firms are approaching innovation "in an opportunistic and ad hoc manner." That finding is now several months old and there's no evidence the picture has materially improved.

The pattern is consistent: firms are interested in AI, they're investing in tools, but the structural barriers - training, leadership capability, change management, and risk culture - aren't being addressed at the same pace. The tools are getting cheaper and more accessible. But access alone doesn't produce adoption.

Meanwhile, Northern Indiana's construction industry is actively recruiting school-leavers by positioning construction trades as "AI-resistant" careers. The data centre build-out is cited as a key driver of sustained employment. It's a useful counterpoint to automation anxiety - and a reminder that construction trades have proved resilient even as AI reshapes adjacent sectors. A similar talent attraction narrative is beginning to emerge in the UK, though less explicitly.