AUS
NZ
UK
CA
USA
IND
Highlights
The AI-infrastructure buildout is creating demand for power generation, grid equipment and precision engineering.
Rolls-Royce's small modular reactor ambitions place it at the intersection of energy and computing.
Names from Segro to BT illustrate how broadly the data centre theme reaches across the London market.
Every technological revolution has its picks and shovels, and the artificial intelligence boom is no different. While the headlines belong to chipmakers and software giants — most of them listed far from London — the physical reality of AI is built from concrete, copper, cooling systems, turbines and transformers. That is industrial territory, and it is increasingly clear that a meaningful slice of the spending wave is flowing towards UK-listed companies. On a day when risk-off sentiment has the broader market near multi-week lows, the durability of this theme is one of the questions occupying London investors.
The core insight is simple: artificial intelligence runs on electricity, and electricity runs on infrastructure. Data centres require enormous and reliable power supplies, sophisticated cooling, robust connectivity and vast sheds of land near transmission capacity. Each of those needs maps onto a set of London-listed businesses, some obvious and some less so. Together they form what might be called the UK's AI-infrastructure supply chain — a collection of industrial and power-linked names whose order books are being reshaped by the buildout.
Why is Rolls-Royce at the centre of the story?
Rolls-Royce (LSE:RR.) is best known for the engines that hang beneath wide-body aircraft, but its relevance to the AI era runs through its power businesses. The group's Power Systems division supplies generation and backup solutions of the kind data centres depend upon, and demand for mission-critical power has become a significant growth engine. More strategically still, its small modular reactor programme positions the company for a future in which hyperscale computing campuses seek dedicated, carbon-free baseload power. Governments and technology companies alike have shown growing interest in nuclear options for exactly this purpose, and Rolls-Royce's selection for national programmes has bolstered the investment case beyond its aerospace recovery.
The wider engineering bench benefits too. IMI (LSE:IMI) supplies flow-control and thermal-management technology relevant to cooling-intensive facilities. Spirax Group (LSE:SPX) brings expertise in steam and thermal energy management. Morgan Advanced Materials (LSE:MGAM) produces speciality materials used across power and electronics applications. None of these is a pure AI play — and investors should be wary of any company marketed as such — but each touches the buildout at points where engineering precision commands healthy margins.
Where do property and connectivity fit in?
The theme stretches well beyond traditional engineering. Segro (LSE:SGRO), the warehouse and industrial property group, controls land and power-connected sites in exactly the locations data centre developers covet, particularly around major European cities. Its ability to convert logistics land into data centre capacity has become an increasingly discussed part of its story. BT Group (LSE:BT.A), meanwhile, owns the fixed-line backbone over which AI services ultimately travel, and the densification of digital infrastructure plays to the value of its network assets.
Power infrastructure specialists complete the picture. National Grid (LSE:NG.), though classified as a utility, sits at the choke point of the entire theme: connection queues for new data centres have become a defining constraint, and the investment required to expand transmission capacity is among the largest capital programmes in the country. Around it, engineering contractors and equipment suppliers — from cable and switchgear specialists to construction groups such as Balfour Beatty (LSE:BBY) — stand to share in the work of physically wiring up the AI economy.
Is the theme already priced in?
That is the debate now playing out across trading desks. Shares with credible AI-infrastructure exposure have generally outperformed this year, and valuation multiples in parts of the industrial sector have expanded well beyond their historical ranges. Bulls argue that the spending commitments announced by technology giants are so large, and stretch so many years into the future, that current order books only hint at what is coming. Sceptics counter that capital expenditure plans can be cut as quickly as they are announced, and that any disappointment in AI monetisation could ripple back through the supply chain with painful speed.
The truth probably sits between the extremes. Infrastructure spending is stickier than sentiment: a data centre half-built will be finished, a grid connection contracted will be delivered. But investors have learned from past cycles — telecoms most famously — that physical buildouts can overshoot demand. The companies best placed are those whose AI exposure supplements an already healthy core business rather than defining it, a description that fits most of the London names involved.
Industrial stocks form one of the largest sector groupings on the London Stock Exchange, covering aerospace and defence, general industrials, electronic and electrical equipment, engineering, construction and materials, and industrial support services under the UK market classification framework. The sector's biggest names, including Rolls-Royce and Melrose Industries (LSE:MRO), are constituents of the FTSE 100, while the FTSE 250 hosts a wide array of specialist engineers and equipment makers. Although adjacent themes pull in companies formally classified elsewhere — property groups such as Segro, utilities such as National Grid, and telecoms such as BT — the industrial core of the AI-infrastructure trade remains firmly within the engineering and capital goods complex, which has long served as a bellwether for global investment cycles.
What could test the trade from here?
Several pressure points deserve attention. Power availability is the most immediate: if grid connections cannot keep pace, project timelines slip and revenue recognition slips with them. Interest rates matter, because infrastructure is financed with debt and higher-for-longer rates squeeze project economics. Geopolitics looms as well — the current Middle East tension has lifted oil prices and injected caution into markets broadly, and a sustained energy shock would raise operating costs across the data centre estate.
Then there is the question hanging over everything: whether artificial intelligence delivers commercial returns commensurate with the staggering sums being invested. UK industrials are, in effect, selling equipment to a gold rush. History is kind to the suppliers of picks and shovels, but only while the rush lasts. For now, the orders keep coming, the power keeps tightening, and London's engine rooms keep humming — a quieter kind of AI story, but one with rivets in it.
