The AI data center power supply chain, Goldman Sachs prefers "gas turbine blades and transformers," presenting a great opportunity for these types of Chinese companies
Goldman Sachs research team pointed out that global AI data centers are facing power bottlenecks, especially in the U.S. market, which presents opportunities for Chinese power supply chain companies with technological and delivery advantages. Goldman Sachs prefers products such as gas turbine blades and transformers, and expects the compound annual growth rate of sales for Chinese companies to reach 23% from 2025 to 2030. The production capacity of U.S. power equipment can only meet 40% of local demand, and the power supply shortage issue is even more severe
The global wave of AI data center construction is facing a serious "power bottleneck," especially in the U.S. market. Goldman Sachs believes this presents a historic opportunity for a group of Chinese power supply chain companies with advantages in technology, cost, and delivery speed.
According to news from the trading desk, on January 14, Goldman Sachs' research team published a report stating that the demand for electricity in global AI data centers is surging, but there is a severe shortage of power generation equipment such as gas turbines and grid equipment like transformers, with delivery times extending to several years.
With orders from European and American suppliers extending to 2028, the shorter delivery times of high-quality Chinese suppliers have become a core competitive advantage. Goldman Sachs believes that the pricing premium driven by supply shortages allows Chinese suppliers to sell at prices 10% to 80% higher in overseas markets compared to domestic prices, and even after accounting for increased tariffs and logistics costs, profit margins still improve significantly.
In terms of product category preferences, Goldman Sachs ranks them as follows: gas turbine blades > power transformers > electrical components > UPS/power racks > liquid cooling > server power supplies (PSU).
Goldman Sachs expects that the average sales compound annual growth rate for Chinese companies from 2025 to 2030 could reach 23%, and by 2030, the overseas AI data center market will contribute an average of 23% of revenue, with an average global market share reaching 4%.
Global AI Data Center Capacity Surge, but U.S. Power Bottleneck More Severe
Goldman Sachs points out that the capacity of global AI data centers (AIDC) is surging, but the power bottleneck issue in the U.S. is far more severe than in other countries.
The report predicts that between 2025 and 2030, approximately 73 gigawatts (GW) of data center capacity will be added globally, with the increase mainly concentrated in the U.S. and China.
Despite the surge in demand, the power supply in the U.S. faces severe challenges, with the bottleneck effect being more pronounced than in places like China:
- Severe local supply shortage: The domestic power equipment capacity in the U.S. can only meet about 40% of its local demand, resulting in a significant supply gap.
- Tight grid reserve capacity: Compared to China, the effective power reserve capacity in the U.S. is much tighter. Report data shows that the average power reserve capacity in the U.S. is expected to be around 100 GW from 2025 to 2030, and the shortage situation will further worsen; meanwhile, China's average reserve capacity during the same period is as high as 300 GW, with a more ample system.
- Extremely high demand growth rate: The report cites internal team forecasts that by 2030, the electricity consumption of data centers in the U.S. (including AI and non-AI) will increase by about 175% compared to 2023, far exceeding the overall power demand's average compound annual growth rate of 2.6%.
- Forced to seek non-traditional suppliers: Due to the long delivery times of key equipment (such as gas turbines and transformers) of 3-5 years, U.S. power companies and data center operators are increasingly willing to accept non-traditional suppliers to cope with supply gaps lasting several years. This creates market opportunities for qualified Chinese suppliers that can deliver quickly Globally, the expansion of AI computing power is shifting from "how many GPUs are needed" to "where the power comes from and how quickly it can be delivered."
In this trend, the United States has become the region with the most severe power bottlenecks due to its limited domestic supply capacity, tight grid reserve capacity, and rapidly increasing demand. This makes suppliers that can provide "Time-to-Power" solutions extremely valuable.
Power generators and transformers are in the tightest supply, with the cycle expected to last until 2028-2030
Goldman Sachs emphasizes that power generators and transformers are the most constrained supply chain bottlenecks in the global construction of AI data centers, and this supply shortage cycle is expected to last until 2028-2030.
The report points out that there has been a multi-year supply shortage in power generation (especially gas turbines) and grid transmission (especially transformers). The pace of capacity expansion is far below initial plans, while demand has surged, leading to a persistent supply-demand gap.
Based on technological barriers and order conditions, the report assesses that the supply tightness of gas turbines (especially blades) will last at least until 2028-2030.
Gas turbine blades must withstand extreme temperatures and physical pressures, involving single crystal high-temperature alloy materials, special coatings, complex aerodynamic designs, and intensive front-end manufacturing processes (such as wax tree manufacturing and assembly). This makes it one of the components with the highest technical barriers and the slowest capacity expansion.
Additionally, due to high customization and labor-intensive characteristics, power transformers face the most severe bottlenecks.
The average age of the power grids in the United States and Europe is 35-40 years, which is already struggling to meet the surging power demand of the AI era, exacerbating the pressure for transformer replacement and upgrades.
Each transformer needs to be custom-designed based on specific impedance, cooling, tap changers, overload, and seismic standards. This "one design per unit" model leads to production being highly dependent on skilled labor and makes it difficult to rapidly expand capacity through capital expenditure.
Technological transformation towards the "great migration" to 800V DC architecture
As the power density of AI cabinets breaks through 100kW and approaches 1MW, traditional copper wire transmission is nearing its physical limits.
To accommodate the massive current, the electrical architecture of data centers is undergoing a dramatic change: transitioning from traditional alternating current (AC) to 800V direct current (DC) architecture. This shift can save approximately 5-15% in energy consumption.
Driven by capacity construction and technological upgrades, Goldman Sachs expects that the power products of AI data centers within the coverage of China's industrial technology will maintain a compound annual growth rate of 39% before 2030.
The research report emphasizes that this will greatly benefit component suppliers with DC technology reserves, especially leaders in the relay and capacitor fields.
Faced with a 3-5 year wait time for key components, this presents a unique opportunity for Chinese suppliers to capture overflow demand Goldman Sachs believes that the core competitive advantage of Chinese suppliers lies not only in cost but also in shorter delivery times and competitive products. Specifically:
Delivery Speed: This is a decisive factor for Chinese suppliers to seize opportunities. Faced with waiting times of up to several years, end customers prioritize "delivery speed" over historical supplier preferences. Chinese suppliers that can provide reliable products and significantly shorten delivery cycles will be favored by the market.
Product Competitiveness: The report emphasizes that opportunities are only available to a few leading Chinese companies that meet international standards in technological strength, quality control, and scale production capacity. These companies typically have expertise in high voltage direct current (HVDC) technology, high-density power conversion, and their products can meet the stringent technical standards and rapidly expanding demands of AI data centers.
Analysis suggests that market entry based on speed and product competitiveness can translate into significant financial gains.
Compared to domestic sales, the premium range for key product categories from Chinese companies can reach 10% to 80%. Although they need to bear additional tariffs and logistics costs, the substantial premium space still allows export businesses to achieve significant gross margin expansion, bringing high visibility in profitability for suppliers.
Goldman Sachs analyzed Chinese suppliers with global competitiveness that can seize opportunities, focusing on the following six product categories:
- Gas Turbine Blades: With high-end manufacturing technology barriers in areas such as single crystal high-temperature alloys, special coatings, and complex aerodynamic designs, these products are among the most critical and technically challenging segments of the global supply chain.
- Power Transformers: Due to high customization, labor intensity, and lengthy certification cycles, these products represent the most severe bottleneck on the grid side. Chinese suppliers have a key advantage in meeting overseas demand due to shorter delivery cycles.
- HVDC (High Voltage Direct Current) Relays: As a key component for data centers upgrading to 800V high voltage direct current architecture, suppliers with advanced HVDC expertise and technical capabilities hold high visibility in global market share in this field.
- Aluminum Electrolytic Capacitors and Supercapacitors: These electrical components are fundamental elements in data center power systems (such as UPS, rectifiers, power factor correction, etc.) and respond quickly to the pull demand from AI data centers. Supercapacitors, in particular, serve as accelerators for peak load regulation in the medium to long term.
- Uninterruptible Power Supply / DC Power Cabinet Systems: Through the ODM (Original Design Manufacturer) model, they can provide high-power supply solutions for overseas markets (especially the United States) and are actively validating and developing next-generation 800V DC architecture products.
- Liquid Cooling Solutions: As the power density of AI servers rises sharply, the demand for efficient liquid cooling systems (including cold plates, manifolds, CDU/RPU, etc.) for chips such as GPUs/ASICs has significantly increased, providing structural growth space and value enhancement for these products
The above wonderful content comes from the Wind Trading Platform.
For more detailed interpretations, including real-time analysis and frontline research, please join the【 **Wind Trading Platform ▪ Annual Membership**】
Risk Warning and Disclaimer
The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at your own risk.
