Irish building systems group Kingspan Group has raised its full-year profit outlook as data centre construction becomes an increasingly important source of growth for a company better known for insulation and building products.

The group now expects trading profit of about €1.13 billion for 2026, around 18% higher than last year and above the €1.05 billion guidance it gave in April. First-half revenue reached €4.86 billion, up 8% year on year, while trading profit increased 10% to €487.2 million despite currency headwinds and costs connected with the abandoned flotation of its ADVNSYS data centre infrastructure business.

ADVNSYS delivered the strongest growth inside the group, with first-half sales rising 34% and both order intake and backlog increasing by more than 100% compared with the same period last year. The division supplies cooling, ventilation, airflow management, and other infrastructure used in high-density data centres, giving Kingspan exposure to an AI investment cycle that extends well beyond processors and software.

Kingspan also expects group revenue to pass €10 billion for the first time, while chief executive Gene Murtagh told analysts that trading profit of €1.3 billion should be achievable in 2027. Those forecasts still depend on performance across the wider group, but the data centre operation is becoming large enough to alter Kingspan’s overall earnings profile.

AI spending spreads into industrial supply chains

As cloud operators and AI companies commit billions to computing capacity, that investment moves through layers of physical infrastructure that receive less attention than semiconductors but are no less necessary once a project moves from procurement into construction. Cooling systems, electrical equipment, ventilation, power distribution, building envelopes, and specialist engineering all become part of the same capital spending chain.

Kingspan’s results sit alongside evidence from other European industrial groups that AI infrastructure demand is becoming visible in conventional order books. Siemens Energy has reported increased demand connected with data centres, while Schneider Electric has also been seeing the physical pull of AI investment through electrical and cooling infrastructure.

That broadening exposure changes how the current data centre cycle should be understood because a significant portion of spending is reaching companies whose revenues were historically tied more closely to construction, industrial investment, and commercial property. A large AI campus may be described as a technology project, but much of its capital budget ultimately pays for buildings, power systems, mechanical engineering, thermal management, and equipment expected to operate reliably for years.

ADVNSYS therefore gives Kingspan an unusual position inside that supply chain, sitting alongside its more traditional building products operations while serving a market with very different demand dynamics. Growth in the division can offset weaker conditions elsewhere, although it also makes group performance more exposed to the timing of large data centre projects and to whether hyperscalers, cloud companies, and developers maintain current investment levels.

Capacity brings a different set of constraints

Strong orders do not guarantee equally rapid delivery because European data centre development is running into physical bottlenecks even where capital is plentiful. Power connections, grid reinforcement, planning, water use, construction capacity, and specialist equipment are influencing where projects can proceed and how quickly completed computing capacity can enter service.

Kingspan invested €233.6 million across the group during the first half, with roughly two thirds directed towards capital expenditure, as it expanded manufacturing capacity in several international markets. It also paused its previously announced €650 million share buyback programme, preserving greater flexibility for acquisitions and investment while opportunities remain unusually large in parts of the infrastructure business.

That choice illustrates the capital allocation problem facing established industrial companies whose end markets are being changed by AI. Returning cash to shareholders competes with the opportunity to expand manufacturing, acquire specialist capabilities, or build capacity for a market growing faster than traditional construction cycles, although investments made during exceptional demand still have to earn returns once the current expansion moderates.

Kingspan’s first-half numbers provide unusually clear evidence of how far the technology investment cycle has moved into the industrial economy. If ADVNSYS continues growing at anything close to its current rate, data centre infrastructure will become less of an adjacent activity for the group and more of a determinant of its wider growth.