Energy Storage System

Investment Pitch:

The Next Wave of AI Growth — Power Flexibility and BESS

  1. Market Overview: A US$1.5 Trillion Thematic Shift

  • Evolution of Investment Trends: AI power demand has catalyzed a rapid re-rating of the global power value chain, creating over US$1.5 trillion in market value in the past two years. 

  • The Next Growth Core: Capital has rotated from nuclear power to grid infrastructure, backup generators, and most recently to fuel cells. Now, Energy Storage Systems (ESS) are emerging as the latest and most critical leg of value creation. 

As AI applications shift from model training to "Inference," the profile of electricity consumption is undergoing a fundamental change: 

  • Inference-Driven Volatility: Unlike predictable training workloads, inference is tied to real-time end-user activity, creating massive instantaneous power spikes and high-frequency noise.

  • Flexibility as a Necessity: Traditional grids and thermal generation are not designed to absorb these rapid, unpredictable swings. BESS, with its millisecond response speeds, is the only efficient solution for managing AI inference volatility. 

2. Core Drivers: From "Base Load" to "Peak Shaving"

  • Infrastructure Deferral Value: BESS acts as a capital efficiency lever, allowing for the deferral or downsizing of expensive, irreversible grid infrastructure investments. Case studies show that deferring a US$10 million investment for five years can generate approximately US$1.15 million in net deferral value. 

  • LCOE Parity: In multiple regions, solar-integrated ESS costs are already competitive with gas peaking plants; in China, they have even reached parity with coal-fired generation. 

  • Spatiotemporal Flexibility: BESS is modular and location-agnostic, allowing deployment directly at load centers like data centers to bypass grid interconnection wait times that can exceed five years. 

3. Advantages: Not Just "Batteries" but "Power Inventory"

4. Market Opportunity & Tech Advancement: The Sodium Era

  • Demand Surge: Global annual incremental ESS deployment for data centers is projected to reach 321GWh by 2030, representing a doubling of current total ESS demand. 

  • Cost-Down Dividends: The commercialization of Sodium-ion batteries is set to further reduce costs. At a production scale of 100GWh, sodium-ion prices could be 30% lower than LFP, while offering superior safety and low-temperature performance.

  • Investing in BESS is not merely an investment in storage hardware; it is an investment in the "Balance Sheet Buffer" of AI infrastructure. By utilizing "Power Inventory," the system can handle structural demand shocks without the risks associated with capital-intensive overbuilding.

5. Investment Value Summary: The Guardian of AI Infrastructure

Significance for Investors:

Growth Certainty:

The share of AI inference is expected to reach 45-50% by 2030, causing peak-shaving demand to surge accordingly.

The bottleneck for AI data center development has shifted from capital availability to "power lead times" and "peak load management". BESS represents the most economical and mature technological pathway currently available to alleviate power shortages and enhance data center uptime.

Policy Tailwinds:

The 30% Investment Tax Credit (ITC) in the U.S. and capacity power price subsidies in China significantly boost project IRRs (reaching as high as 22-36%).

Capital Efficiency:

Replacing "Capacity (Power Plants)" with "Inventory (Energy Storage)" reduces the risk of systemic overbuilding.