The AI Power Boom Could Create a Repair Business for Old Turbines

The artificial intelligence boom has created an unusual problem. A developer can secure funding for a data centre and still wait years for its power equipment.

On 9 November 2026, Reuters reported that developers were seeking used gas turbines and repurposed engines as new-machine supply lagged demand. It cited worldwide turbine orders above 100 gigawatts against estimated annual production capacity of 60–70 gigawatts.

A separate order shows why delivery time matters. Wärtsilä announced a 282-megawatt US data-centre power project on 5 October. It booked the order in the third quarter of 2026, with engine delivery planned for 2028 and operation expected in 2029. These are new engines, not refurbished turbines. Reuters’ 10 November report on Venture Global adds wider context about companies expanding into AI power infrastructure, rather than evidence of repair contracts.

We think the connection is a possible opening for firms that can make existing equipment usable before a replacement arrives.

This could create inspection and condition-assessment fees for qualified engineers. A buyer needs to understand a machine’s history, damage and remaining operating life before deciding whether it is worth restoring.

A second opening is refurbishment and replacement-parts work. A specialist could earn money by restoring a suitable machine to an agreed standard. The commercial advantage depends on the full installed cost, delivery time, permits and reliability, not the purchase price alone.

A third opportunity is long-term maintenance and monitoring contracts. If a restored machine enters service, its owner may pay for regular checks, replacement parts and a clear repair-response plan. That work could continue after installation, but recurring revenue is not yet established by the evidence reviewed here.

Listed research examples include Sulzer (SIX: SUN), Siemens Energy (Xetra: ENR), GE Vernova (NYSE: GEV), Wärtsilä (Nasdaq Helsinki: WRT1V) and Caterpillar (NYSE: CAT). Reuters links Sulzer to used-turbine enquiries, Siemens Energy and GE Vernova to turbine manufacturing, and Caterpillar to its Solar Turbines business. Wärtsilä supplies the new-engine example above. These are different exposures, not five confirmed winners from refurbishment.

Our prediction: over the next 6–18 months, paid inspection, repair and support work could grow if developers can install suitable used equipment sooner and at an acceptable total cost. A funded repair order would be more useful evidence for this thesis than another broad AI investment announcement. A construction slowdown, released equipment reservations or faster new deliveries could reduce the opening.

Commercial test

A qualified service firm should start with an authorised equipment owner. Obtain the machine’s maintenance records, intended operating conditions and inspection scope before quoting. Check technical suitability, approvals, warranty obligations and who pays for an unexpected defect.

For illustration only, suppose a refurbishment contract brings in US$400,000 and parts, labour, testing and transport cost US$300,000. The US$100,000 difference is before overhead, financing, insurance, tax and unplanned repairs. One major defect could remove it. These are hypothetical planning numbers, not observed prices, customer budgets or evidence of profit.

For a stock investment, establish the share price, valuation, service margins, order conversion, balance-sheet strength and the portion of earnings linked to this work. Those checks are not complete here.

Act when: a named operator offers a funded, technically suitable refurbishment scope, or a supplier discloses meaningful repair-and-service bookings.

Stop when: repair costs, equipment condition, permit delays or new-machine availability remove the used equipment’s cost or timing advantage.

Report this passage