Somewhere in the northwest corner of Saudi Arabia, the world’s biggest construction project is underway. NEOM’s THE LINE, a 170-kilometre mirrored city designed to house 9 million people, is already the world’s most ambitious test of what AI can do at construction scale. And very few people in UK construction are paying close attention to it.
They should be.
The overall NEOM programme is projected to cost £1.2 trillion ($1.5 trillion). That is not a misprint. It is the largest single capital programme in human history. And it is being built, right now, using AI-driven logistics, autonomous construction equipment, robotic rebar assembly, and a digital infrastructure backbone that most UK firms have not even started thinking about.
For a QS firm, a Tier 1 contractor, or an FM operator looking at where AI in construction is actually heading, this is the live demonstration you need to understand.
What NEOM Is Actually Deploying
Let’s be specific. This is not vaporware.
In December 2024, NEOM’s Investment Fund signed an agreement with GMT Robotics, a Copenhagen-based firm, to deploy robotic rebar cage assembly at scale. Their systems reduce on-site workforce requirements by an average of 90% through offsite prefabrication. See: neom.com/newsroom/neom-investment-fund-ventures-into-automated-robotic-technology
At Oxagon, NEOM’s industrial city on the Red Sea coast, the Port of NEOM welcomed its first fully automated, remote-controlled cranes in June 2025. These are the first automated Ship-To-Shore and Electric Rubber-Tyred Gantry cranes in Saudi Arabia. Terminal 1 opens in 2026 with full horizontal transport automation. See: neom.com/newsroom/port-of-neom-strengthens-role-in-global-supply-chain-connectivity
In February 2025, DataVolt signed a landmark agreement to build the region’s first net-zero AI factory campus at Oxagon. A 1.5-gigawatt facility, spanning 350,000 square metres, powered entirely by renewable energy. Phase one investment: $5 billion (£3.9 billion). Operational by 2028. See: neom.com/newsroom/datavolt-signs-agreement-with-neom.
The contrast with UK construction is stark. According to CECA’s May 2025 report, 85.6% of UK construction businesses reported not using AI as of March 2025, compared to almost 75% of businesses in other sectors. The sector lags the broader economy, despite AI being described by CECA as something that could “set UK construction on its next major leap in terms of performance.”
The tools exist. The gap is in adoption, and in governance.
The Governance Question Nobody Is Asking
“AI adoption in construction lags other sectors. Skills shortages, data quality issues, and integration with existing systems remain significant barriers, while questions around ownership of data for training AI models persist.”
Here is the thing NEOM AI construction reveals that UK firms would rather not think about: when you deploy AI at this scale, governance is not optional. It is an engineering requirement.
NEOM is running AI logistics across a 26,500 km² site. It is using autonomous systems to coordinate thousands of vehicles, cranes, and earthworks machines. It is training models that will eventually manage an entire city’s infrastructure. Someone has to decide: who owns the data? What happens when the AI makes a wrong call on a safety-critical system? Who carries the liability?
These are not hypothetical questions for NEOM and AI-Powered Construction. They are live engineering governance problems being solved right now. And here is the uncomfortable truth for UK construction: if a programme with this level of resource and technical ambition needs AI governance frameworks baked into the architecture from day one, what does that say about a Tier 2 contractor deploying a scheduling AI tool on a hospital project in Leeds?
It says the same rules apply. At different scale, yes. But the same questions about data ownership, liability, human oversight, and model accountability are sitting there, unanswered, in every firm that has plugged in an AI tool without a governance framework to back it up.
CECA’s February 2026 update is clear on this: “In safety-critical contexts, including design and site operations, the principle that AI must not displace or dilute human responsibility is increasingly reflected in regulatory thinking, which is emphasising human oversight, validation, and accountability.” That is the direction of travel. Firms that are not building governance into their AI adoption now will be scrambling to retrofit it when the regulations land.
The Practical Takeaway for UK Firms
You do not need a £1.2 trillion programme to start learning from NEOM. Three things are transferable right now.
First: the greenfield advantage. NEOM is building digital infrastructure from scratch, which means it can embed AI-Powered Construction and automation into every layer of its logistics, procurement, and site management without fighting legacy systems. UK firms cannot do that for every project, but they can insist on it for new digital tool deployments. Do not retrofit. Design for AI from the start.
Second: robotic prefabrication is not futuristic. It is happening now, at cost, at scale. GMT Robotics is a European company. The technology NEOM is deploying is available to UK contractors. The question is not whether robotic rebar and automated cranes will reach UK sites, it is how quickly, and whether your firm is ready to manage them safely.
Third: NEOM’s AI governance problems are your AI governance problems. The scale is different. The questions are the same. If you are deploying AI on site, for scheduling, safety monitoring, cost forecasting, or design, you need a framework that covers data ownership, model accountability, and human oversight. Not eventually. Now.
