Theme & Background
This chapter discusses the accelerating growth trend in global electricity demand driven by artificial intelligence, electrification, and advanced manufacturing. The report notes that U.S. electricity demand has remained largely flat over the past 15 years, but this backdrop is rapidly changing, with electricity transitioning from a secondary input to the core of investment discussions and digital growth narratives.
Core Thesis
The author argues that electricity demand growth is persistent and structural, rather than cyclical. Companies involved in power generation, electrical equipment, grid infrastructure, industrial automation, and transmission networks stand to benefit. A counterintuitive judgment is that AI investment opportunities lie not only in AI labs and hyperscale cloud providers but also in the bottleneck segments supporting this infrastructure (such as transmission, transformers, and on-site power generation).
Key Arguments & Data
- Data Center Electricity Demand: Global data center electricity demand is expected to roughly double from approximately 415 terawatt-hours in 2024 to about 945 terawatt-hours by 2030.
- U.S. AI Power Generation Demand: AI may require up to 90 gigawatts of incremental generation capacity over the next five years.
- Transmission Bottlenecks: Lead times for large substations and generator step-up transformers range from approximately 80 to 210 weeks, reflecting sustained demand exceeding supply.
- Labor Scarcity: Skilled tradespeople require years of training and are difficult to replicate, representing a core bottleneck in infrastructure construction.
- Data Center Growth Contribution: Over the next five years, data centers are expected to account for approximately 55% of the growth in U.S. electricity demand.
| Metric |
Data |
| Global data center electricity demand (2024) |
Approximately 415 TWh |
| Global data center electricity demand (2030 forecast) |
Approximately 945 TWh |
| Incremental U.S. AI generation capacity demand over five years |
Up to 90 GW |
| Large transformer lead times |
Approximately 80–210 weeks |
| Data centers' share of U.S. electricity demand growth over five years |
Approximately 55% |
Companies/Assets Covered
- Quanta Services: A key player in transmission infrastructure. The company employs approximately 70,000 skilled tradespeople and reported a record backlog of $44 billion in 2025. The report is bullish, arguing that its engineering capabilities, project management, and scarce labor force are increasingly valued in the transmission expansion.
- Hyosung Heavy Industries: A transformer manufacturer. Its heavy industry division produces large power transformers and gas-insulated switchgear, representing one of the most critical bottlenecks in transmission construction. The report is bullish, citing triple demand drivers (aging grid replacement, new loads from AI and electrification, and long-distance transmission needs), the highest value in high-voltage transformers (765 kV), competitive advantages from U.S. domestic production capacity (Memphis plant), and potential for further margin improvement through geographic and product mix optimization and price inflation.
- Siemens Energy: A gas turbine manufacturer. Its gas turbines are increasingly used for on-site data center power generation, as grid connection bottlenecks cause project delays. The share of orders from data centers has risen from negligible a few years ago to over 20% currently. The report is bullish, arguing that as electricity demand grows faster than grid expansion, the value of dispatchable generation (such as gas turbines) rises, particularly for large AI data centers requiring rapid power supply.
Investment Implications
Investors should focus on bottleneck segments in power infrastructure rather than solely on the generation side. Specific directions include:
- Transmission Network Construction: Such as Quanta Services, benefiting from labor scarcity and growing backlogs.
- Transformer Manufacturing: Such as Hyosung Heavy Industries, benefiting from supply tightness and product upgrades.
- On-Site Power Generation Equipment: Such as Siemens Energy, benefiting from grid connection bottlenecks and AI data centers' demand for rapid power supply.
These areas may offer more durable pricing power and growth potential due to constrained capacity expansion, labor scarcity, and inelastic demand.
Theme and Background
This chapter focuses on alternative solutions for providing on-site power to AI infrastructure and summarizes the investment opportunities within the broader electricity theme. The report notes that grid interconnection delays and turbine supply constraints make modular, rapidly deployable on-site power generation solutions more attractive. The author argues that electricity has become a critical constraint on digital growth, industrial expansion, and economic resilience, and that the investment horizon should extend beyond chips and models to the physical infrastructure underpinning AI.
Core Views
- Speed as a Competitive Advantage: Against the backdrop of lagging grid construction and constrained supply of traditional power generation equipment (e.g., turbines), on-site power solutions that can be deployed quickly (such as Bloom Energy's fuel cells) are being upgraded from backup power to primary power sources for data centers.
- Opportunities Broader Than They Appear: Investment opportunities in the AI electricity theme are not limited to the power generation side but span multiple bottleneck areas, including labor (Quanta), transformers (Hyosung), power generation and grid modernization (Siemens Energy), and on-site modular solutions (Bloom Energy).
- Not a Style Drift: Although Sands Capital has traditionally not invested in power equipment and engineering, procurement, and construction (EPC) in developed markets, the structural shift in electricity demand allows select companies to leverage their existing market positions to capture a larger share of industry profits.
Key Arguments and Data
- Bloom Energy's Positioning: Its fuel cell systems are modular and designed for scenarios requiring "always-on" operation and rapid deployment. Customers are using them as baseload power for data centers, not merely as backup power.
- Four Bottlenecks Corresponding to Four Companies: The report maps four companies to different bottlenecks in AI power supply:
- Quanta Services: Labor and execution backbone
- Hyosung Heavy Industries: Transformer bottleneck
- Siemens Energy: Dispatchable power generation and grid modernization
- Bloom Energy: Faster, more modular on-site power solutions
- Screening Criteria: Sands Capital emphasizes that these companies were selected based on their largest growth opportunities, most innovative leading products and services, and most robust business models, enabling them to benefit more fully from this long-term trend.
Companies/Assets Involved
| Company |
Role and Key Data |
Bullish/Bearish |
| Bloom Energy |
Modular fuel cell systems offering a "speed-to-power" competitive advantage. Used for data center baseload, not just backup. Technology-driven, with a wider range of outcomes. |
Bullish (but higher risk) |
| Quanta Services |
Represents the labor and execution backbone of power construction. |
Bullish |
| Hyosung Heavy Industries |
Represents the transformer bottleneck in transmission systems. |
Bullish |
| Siemens Energy |
Represents dispatchable power generation and grid modernization on a global scale. |
Bullish |
Investment Implications
- Focus on the "Speed" Premium: In an environment constrained by grid and traditional equipment supply, companies offering rapidly deployable on-site power solutions (such as Bloom Energy) may command a higher valuation premium, despite higher technology risk.
- Broaden the Investment Horizon: The next phase of AI growth depends not only on better models or faster chips but also on companies building the underlying physical systems. Investors should systematically examine each bottleneck in the power infrastructure (labor, transformers, generation equipment, modular solutions) rather than focusing solely on the most obvious beneficiaries.
- Selection is Key: Not all electricity-related companies are worth investing in. Focus should be on those with the largest growth opportunities, the most innovative products, and the most robust business models, as they are best positioned to capture the profits from this structural shift.