AI data centers are accelerating the buildout of on-site power, with demand for natural gas power turbines growing faster than utilities can expand electricity generation and grid connections. Industry estimates cited in a recent analysis by BloombergNEF and PwC point to a multi-year imbalance that is increasingly pushing turbine makers and heavy-equipment suppliers into the spotlight.
That surge is also colliding with supply constraints. Wood Mackenzie said the per-kilowatt cost of gas-powered turbines could be 195% higher by the end of next year than it was in 2019, creating a pricing environment that benefits companies positioned to deliver hardware and related systems to data center operators.
Key takeaways
-
Natural gas turbine demand linked to AI data centers is outpacing utility electricity capacity expansion, with more projects moving toward self-sufficient generation.
-
According to BloombergNEF, nearly 100 data centers have or are building on-site natural gas turbine power infrastructure, supported by PwC’s forecast for natural gas consumption to more than quintuple by 2035.
-
Cost pressures are rising: Wood Mackenzie estimated turbine costs per kilowatt could be as much as 195% higher by the end of next year versus 2019.
-
The implication for investors is that turbine and on-site power suppliers—especially those with strong order pipelines—may see sustained revenue support even as procurement becomes more expensive.
What’s driving AI data centers toward natural gas turbines
Utilities generally cannot deliver the scale of power hookups that AI operators want on the timelines being pursued for new data center campuses. Natural gas turbines offer an alternative: large onsite units that can produce electricity directly, reducing reliance on grid expansion and potentially improving scheduling certainty.
BloombergNEF estimates there are nearly 100 data centers either operating with, or in the process of building, on-site natural gas turbine power infrastructure. PwC also projected that AI-related natural gas consumption could more than quintuple by 2035, with power turbines playing a central role in meeting that demand.
Why turbine supply and pricing matter to markets
The central issue for operators is that demand is expanding faster than available turbine capacity and manufacturing supply chains can keep up. Wood Mackenzie, in commentary earlier this year, said the per-kilowatt cost of gas-powered turbines could rise to 195% above 2019 levels by the end of next year.
For investors, the market implication is twofold. First, higher prices can support revenue growth for turbine manufacturers and component suppliers with sufficient backlog. Second, cost inflation can increase uncertainty for data center project economics and could affect how quickly operators can finalize new installations—though the analysis argues that demand will continue to exceed supply “well into the foreseeable future.”
Companies positioned for the turbine cycle
GE Vernova
GE Vernova is a leading name in gas turbine power equipment, with the turbine segment currently described as its leading profit center for the “foreseeable future.” The company’s most recent quarter showed organic revenue growth of 12%, with the power division—described as including gas turbines—contributing 14% growth.
Order momentum is also a key part of the thesis. The analysis said total orders for the gas turbine unit jumped 134% year over year in Q2, increasing backlog by $13 billion to $176 billion. The backlog is presented as large relative to annualized revenue, with expectations that it can keep growing as data center builds continue.
Siemens Energy
In Europe, Siemens Energy is highlighted as a counterpart to GE Vernova for North American demand and a supplier with growing involvement in gas turbine buildouts. According to the analysis, Siemens Energy delivered 6 gigawatts’ worth of gas-powered turbines during the quarter, while receiving 15 gigawatts of new orders, lifting its backlog to 69 gigawatts of gas-power equipment.
The report also cited revenue growth of 18.5% year over year in the most recent quarter, attributing a large portion of the increase to AI data center demand.
Mitsubishi Heavy Industries
Mitsubishi Heavy Industries is also identified as an important turbine supplier, with the analysis noting that it reported revenue growth of 13.3% in its most recently completed quarter. As with Siemens Energy and GE Vernova, the argument is that the company is adding capacity to meet demand that exceeds what can be delivered immediately.
The article further emphasized that, while Mitsubishi’s and Siemens Energy’s U.S. listings are described as over-the-counter, the underlying businesses have conventional exchange listings in their home countries.
Additional beneficiaries in the broader on-site power supply chain
Beyond pure-play turbine makers, the analysis pointed to heavy equipment and power generation equipment providers that can participate in the onsite power buildout as data centers scale.
Caterpillar
Caterpillar is described as an additional beneficiary because some AI data center operators are using its diesel-powered generators as a power source. The analysis cited Microsoft’s planned Monarch Compute Campus in West Virginia, stating that the facility will initially rely on Caterpillar’s G3500-series natural gas generators before transitioning to gas-powered turbines made through Caterpillar’s wholly owned subsidiary Solar Turbines.
The piece also said a large share of Caterpillar’s 24% year-over-year sales growth last year was driven by data center demand.
Woodward
Woodward was included as a company serving the on-site power production market. The analysis said that natural gas power turbines are not currently described as a major profit center, but it suggested Woodward could benefit from the broader demand for components and systems used in onsite generation.
For investors, the report framed GE Vernova as the most straightforward option among the turbine-focused names, mainly due to the level of reported segment activity and backlog support discussed in the analysis.
Bigger picture: what to watch next for the AI power buildout
With utilities still constrained by generation additions and grid timelines, investors will likely keep focusing on how quickly turbine manufacturers can convert order flow into deliveries and how pricing translates into margins. The next signals to monitor include major turbine order announcements, backlog revisions, and updates on data center commissioning schedules. On the macro side, energy price volatility and industrial capacity constraints—both of which can influence project timelines and procurement decisions—remain important variables as the AI build cycle continues.
