Summary
- Opna warns that critical power equipment shortages could become a major barrier to Scottish data centre growth.
- The constraint covers transformers, cables, switchgear, and access to manufacturer production slots.
- Network operators’ large framework agreements may leave individual data centre projects competing from the back of long order books.
Opna has warned that shortages of critical power equipment could delay Scotland’s planned data centre expansion, turning transformers, switchgear, cables, and manufacturer access into a practical constraint on AI infrastructure growth.
The London-based critical power equipment market infrastructure company argues that Scotland’s grid investment and data centre pipeline are competing for the same electrical supply chain. Network operators, with bulk purchasing power and long-term framework agreements, are more likely to secure manufacturer capacity than individual data centre projects placing one-off orders.
The warning follows concern that an £8.2bn AI data centre project in Lanarkshire, led by CoreWeave and DataVita, may struggle to meet its original 2030 operational target. The constraint is not limited to one project. Scotland is attracting interest because of renewable energy resources, available land, and the UK’s AI infrastructure agenda, but those advantages still have to pass through the physical bottleneck of grid connection equipment.
DCNN reported Opna’s comments alongside figures showing the scale of the pressure. Scottish wind farms received about £343m in curtailment payments in 2025, while Wood Mackenzie has reported average transformer lead times of 128 weeks, with some orders extending beyond four years and prices rising sharply since 2019.
The bottleneck behind the bottleneck
Data centre power discussions often stop at grid capacity. Capacity only becomes useful when the equipment, engineering, permissions, and commissioning path exist to deliver it. Transformers, switchgear, protection systems, cabling, and substations are not generic commodities that can be bought at short notice once a project reaches financial close.
That creates a difficult sequencing problem for developers. Ordering long-lead electrical equipment too late risks delaying energisation. Ordering too early can tie up capital and expose the project to changes in design, planning, customer demand, or connection terms. Opna’s proposition is built around that gap: sourcing, financing, and delivery models that reserve or access equipment capacity before projects are ready to draw it down.
The issue is acute for data centres because electrical infrastructure is not peripheral to the business model. Megawatts are the product. A data centre without its grid connection, substations, and internal electrical distribution is a building shell with a cooling problem. Hyperscale and AI facilities intensify that dependency because their loads are larger, denser, and more sensitive to delays in high-voltage equipment.
Scotland’s grid investment programme adds another layer. SP Energy Networks has begun a £12bn programme of upgrades across central and southern Scotland, including new substations and a supply-chain framework worth up to £5.4bn over 10 years. SSEN Transmission is investing at least £22bn in northern Scotland by 2031 and has announced a £7.4bn supply-chain framework. Those programmes are needed to unlock capacity, but they also pull on the same equipment categories as data centre developers.
Power-rich does not mean plug-ready
Scotland’s appeal to data centre developers is clear. The country has substantial renewable generation, policy interest in AI and digital infrastructure, cooler ambient conditions, and industrial sites that may be more available than in congested southern markets. Power-rich regions, however, are not automatically plug-ready regions.
Wind curtailment illustrates the difference. Available generation does not mean electricity can be moved, transformed, and supplied exactly where and when a large data centre needs firm capacity. Transmission constraints, distribution reinforcement, connection queues, substation availability, and equipment procurement all shape whether renewable power can become usable data centre load.
The equipment shortage also affects project finance. Lenders and investors increasingly want evidence that grid access is more than a line in a development presentation. They want credible connection milestones, procurement plans, and delivery risk management. If transformer or switchgear lead times are measured in years, power equipment strategy becomes part of underwriting.
For policymakers, the problem cuts both ways. Prioritising network operators in equipment supply is logical because the grid needs reinforcement to serve all users. But if data centre growth depends on the same supply chain, grid upgrades and large-load connections can become mutually congested. A national data centre strategy or AI growth zone policy that ignores equipment deliverability will overstate the speed at which capacity can come online.
The practical response is likely to involve earlier procurement planning, standardised designs, better visibility of order books, framework access for developers, and closer coordination between network operators, data centre customers, equipment manufacturers, and financiers. Scotland’s data centre debate is already politically charged around land, power demand, and community impact. The power warehouse now needs as much attention as the power station.
