Replacing gas turbines, the top choice for AI-powered electricity is a gas generator set?

AI data centers’ demand for power equipment is reshaping the entire generation equipment market landscape. A recent research report by HSBC Bank states that natural gas power generation sets are rapidly replacing gas turbines, becoming the AI data centers’ preferred transitional primary power source, and the situation of supply-demand imbalance will be difficult to ease in the short term.

According to the Tracking Wind Trading Desk, HSBC recently released a report: due to a surge in demand driven by AI, orders for leading natural gas power generation set manufacturers have been booked out through the end of 2027, and equipment prices in 2025 have already risen by 15% to 20%.** The tight supply situation is expected to continue through 2027 to 2028, during which equipment prices are still expected to increase by 10% to 15% each year.**

Gas power generation sets: Why they’ve become the new favorites of AI data centers

The rise of natural gas power generation sets in the primary power market for AI data centers, stems from their clear advantages in three areas: delivery cycle, start-up performance, and flexibility.

First, the delivery cycle of natural gas power generation sets is 1 to 2 years, which is significantly shorter than the 2 to 4 years required for gas turbines. This enables data centers to match their expansion pace more quickly.

Second, in terms of cold start time, natural gas power generation sets need only 30 to 60 seconds, far faster than the 5 to 60 minutes required for gas turbines. This is crucial for AI training scenarios where load fluctuates sharply.

In addition, the single-unit capacity of power generation sets ranges from 2 to 8 megawatts, with a high degree of modularity. They can be expanded as needed, and after future grid interconnection is completed (expected around 2030), they can be flexibly repurposed as backup power.

By comparison, large gas turbines can reach single-unit capacities of 100 to 500 megawatts or more, but they have long deployment cycles and weaker flexibility. They are more suitable for large cloud computing facilities with stable demand—not AI training scenarios with aggressively shifting load characteristics.

Supply-Demand Imbalance: Orders Are Fully Booked, and Prices Keep Climbing

In the high-speed natural gas power generation set segment, Caterpillar and INNIO Jenbacher together account for about 65% of market share, with orders for both companies already filled through the end of 2027. The medium-speed power generation set market is dominated by Wärtsilä and Everllence, with a combined market share of about 75%, also operating at full load.

In terms of pricing, the current total capital expenditure for deploying natural gas power generation sets in data centers is about $1,400 to $1,700 per kilowatt. Of this, the bare power generation set itself is about $600 to $1,000, auxiliary equipment is about $200, and exhaust gas treatment is about $100.

Bare-unit price quotes vary noticeably across major manufacturers: Caterpillar, CMI, and MTU (2 to 8 megawatt models) are about $600 to $650 per kilowatt; Jenbacher is about $750 to $800; Wärtsilä and Everllence (20 to 25 megawatt models) are as high as $800 to $1,000.

The supply-demand balance for natural gas power generation sets is expected to rebalance around 2028 to 2029, when prices may see a fairly significant pullback. By comparison, the supply-demand turning point for diesel power generation sets is expected to arrive earlier, around 2027 to 2028.

Key Beneficiary Stocks: Caterpillar and Weichai Power

HSBC highlights two stocks in the report, each for different reasons.

HSBC views Caterpillar as an “indispensable” supplier in the construction of AI data centers.

The company’s product line includes diesel power generation sets, natural gas power generation sets, and gas turbines. Its market position is strong. In the high-speed natural gas power generation set market, Caterpillar ranks first with about a 35% market share; in the diesel power generation set market, it also holds a leading position.

Weichai Power is seen as an opportunity for potential market share gains. Weichai has three catalysts:

One is a shorter delivery cycle (30 to 60 weeks, versus about 100 weeks for Caterpillar and Cummins); two is that product plans for natural gas power generation sets are scheduled to be launched in the second half of 2026. The 2 to 3 megawatt models are expected to go on the market in June 2026, while the 5 megawatt and 7 megawatt medium-speed models are expected to be released in December 2026; three is that capacity for solid oxide fuel cells (SOFC) continues to ramp up.

The above excellent content comes from the Tracking Wind Trading Desk.

For more detailed analysis, including real-time interpretations, first-line research, and more, please join 【**Tracking Wind Trading Desk ▪ Annual Membership**】

Risk Disclosure and Disclaimer Terms

          

            The market involves risk; investments require caution. This article does not constitute personal investment advice, and it does not take into account any individual user’s special investment objectives, financial conditions, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article align with their specific circumstances. Accordingly, any investment decision is at your own risk, and responsibility is assumed by you.