Data center battery demand rises, but market sentiment cools due to overcapacity and uncertainty around AI capex

Overview

This report focuses on global energy storage, especially the upward revision to battery demand driven by U.S. data centers. The authors believe data centers are becoming more "battery intensive," with applications expanding from short-duration UPS and BBU to longer-duration BESS/ESS, and some hyperscale data center projects reaching 4-8 hours of storage duration. Despite stronger long-term demand, supply expansion is still outpacing demand, and combined with uncertainty around returns on AI data center capex and Korean battery makers' margins, the investment view remains neutral with caution.

Core views

The core view is strong demand, even stronger supply, improving valuations, but entry points still need patience. Rising data center demand drives an upward revision to U.S. battery demand, with ESS becoming the largest incremental source; however, even though the U.S. 2030 capacity forecast is cut from 798GWh to 721GWh, it still exceeds 486GWh of demand, so structural oversupply may persist. Korean battery makers have relative advantages amid U.S. localization, tariffs, and subsidies, and Samsung SDI benefits more from higher ESS mix and lower EV exposure, but the report still maintains Market-Perform.

Analysis framework

The report combines announced U.S. data center projects, data center power capacity forecasts, and penetration and duration assumptions for different battery applications to estimate data center battery demand; it then incorporates this into the framework for total U.S. battery demand, capacity, and utilization to assess supply-demand balance and the impact on company revenue, profit, and valuation. At the company level, DCF, SOTP, and comparable valuation frameworks are used to update target prices and earnings forecasts for LGES, LG Chem, and Samsung SDI.

Methodology notes

• Demand ForecastingData center BESS demand model

  Estimate battery demand based on data center power capacity, BESS penetration, and effective storage duration

  The report assumes U.S. data center power capacity rises from 41GW in 2025 to 103GW in 2030, BESS penetration reaches 100% by 2030, and effective duration increases from 4.0 hours in 2024 to 6.0 hours in 2030, thereby deriving 277GWh of cumulative data center battery installations in 2030.

• Supply-Demand AnalysisU.S. battery capacity and demand balance

  Compare U.S. battery demand, manufacturing capacity, and utilization

  The report includes ESS, data center, and EV demand in its total demand forecast and compares this with U.S. domestic capacity expansion, concluding that structural oversupply may still exist in 2030 even after demand is revised up and capacity forecasts are revised down.

• Valuation methodsDCF

  Discounted cash flow valuation

  LGES uses a 9.6% WACC and 3% terminal growth rate, while Samsung SDI uses a 9.3% WACC and 3% terminal growth rate, with the models including annual free cash flow forecasts through 2050 and terminal value.

• Valuation methodsSOTP

  Sum-of-the-parts valuation

  LG Chem uses the SOTP approach, in which the Energy Solutions business uses DCF, Petrochemicals uses 10x 2025-27 P/E, Advanced Materials uses 2.0x 2025 EV/Sales, and Life Sciences uses 50x 2025-27 P/E.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• LG Energy Solution (373220.KS)

  One of the key beneficiaries of U.S. domestic battery and ESS supply, with the report maintaining Market-Perform and raising the target price to KRW351,000.

  Strengths

  Largest total U.S. capacity scale, expected to reach about 198GWh by end-2026 and 232GWh by end-2027; dedicated ESS capacity is expected to reach 48GWh starting in 2026; benefits from U.S. subsidies and localized demand.

  Weaknesses

  FY26 EPS is cut into negative territory, and margin recovery remains uncertain; the Ohio battery plant suspension leads to lower capacity estimates.

  Comparison

  Compared with Samsung SDI, LGES has larger scale but more pronounced EV exposure and margin pressure; compared with Chinese battery makers, it is more strongly protected by the U.S. tariff wall and subsidies.

  Risks

  Declining IRA tax incentives, reallocation of tax credits with customers or partners, manufacturing cost inflation, and insufficient capacity utilization.

• Samsung SDI (006400.KS)

  A relatively more direct Korean beneficiary of the ESS trend, with the report maintaining Market-Perform and a KRW520,000 target price.

  Strengths

  ESS capacity is expected to reach 30GWh in 2026, accounting for more than half of its U.S. battery manufacturing capacity; EV exposure is relatively lower, and nearly half of battery sales may come from ESS; the long-term target price implies higher upside.

  Weaknesses

  FY26 EPS is cut, and near-term earnings remain under pressure; valuation still depends on whether high-single-digit margins can be delivered.

  Comparison

  The report believes SDI has greater exposure than LGES to the rising ESS trend, but still recommends waiting for a better entry point.

  Risks

  A slowdown in AI data center capex, weaker-than-expected execution of ESS capacity conversion, delays in margin recovery, and industry oversupply pressuring prices.

• LG Chem Ltd (051910.KS)

  Benefits from energy storage demand through its Energy Solutions business and exposure to LGES, with the report maintaining Market-Perform and raising the target price to KRW312,000.

  Strengths

  In the SOTP, Energy Solutions accounts for about 70% of the target price, and upward revisions to energy storage demand support valuation; the target price increase reflects the demand update.

  Weaknesses

  The group structure is more complex, with divisions such as Petrochemicals, Advanced Materials, and Life Sciences diluting pure energy storage beta; FY26 EPS is cut.

  Comparison

  Compared with LGES and SDI, LG Chem is a more diversified chemicals and battery-related asset, with less direct energy storage beta than pure battery makers.

  Risks

  Changes in battery business valuation assumptions, petrochemical cycle volatility, cost inflation, and uncertainty around tax incentives.

Key data

• U.S. data center power capacity2030E 103GWThe report assumes growth from 41GW in 2025 to 103GW in 2030.
• U.S. cumulative data center battery installations2030E 277GWhThis corresponds to the data center battery/GW multiple rising from about 0.3x to 2.7x.
• U.S. annual data center battery demand2030E about 66-69GWhDifferent sections of the report cite 66GWh, 66.4GWh, and 69GWh, all pointing to roughly 10x growth.
• Total U.S. battery demand in 2030486GWhRaised by 74GWh from the previous forecast, implying about 22% CAGR through 2030.
• U.S. battery capacity in 2030721GWhCut from the previous 798GWh, but still above the demand forecast, supporting the structural oversupply view.
• 2030 demand mixESS about 40%, data centers about 14%Downgrades to EV demand are offset by upward revisions to ESS and data center demand.
• Domestic ESS capacityabout 102GWh in 2026, about 122GWh in 2027Driven by LGES, Samsung SDI, SK On, Tesla, and others shifting toward ESS and LFP applications.

Impact & implications

For investors, data centers and ESS raise the long-term demand slope for the battery industry and improve the medium- to long-term risk-reward profile of Korean battery makers; however, continued U.S. domestic capacity expansion may still weigh on utilization and margins, and a near-term equity rerating requires clearer evidence on AI capex sustainability, the pace of EV-to-ESS capacity conversion, proof of margin recovery, and further capacity reduction announcements.

Risks

• If AI and hyperscale data center capex slows due to uncertainty around returns, it could weaken data center battery demand.
• U.S. battery capacity expansion may still outpace demand growth, leading to structural oversupply, project delays, phased commissioning, or cancellations.
• Korean battery makers' margins excluding subsidies are close to break-even, and whether they can turn positive in 2H26 remains debated.
• U.S. EV demand has been revised down by about 25% year to date, which may continue to pressure battery makers' near-term margins.
• Declining IRA tax incentives, reallocation of tax credits, or manufacturing cost inflation may weaken the profitability of companies such as LGES.
• The scale and timing of converting domestic ESS capacity from EV production lines have not been fully disclosed, and execution risk remains high.

What to watch

• Mid-year earnings and capex guidance from hyperscale cloud providers, especially the pace of AI data center buildout.
• Announcements of paused, canceled, delayed battery capacity additions in the U.S., or EV-to-ESS conversions.
• Whether Korean battery makers' 2H26 margins can turn positive as guided.
• The ramp-up progress of domestic U.S. ESS capacity, especially conversion plans at LGES, Samsung SDI, SK On, and Tesla.
• The extent of continued protection for the domestic U.S. battery supply chain from IRA, tariffs, and subsidy policies.
• Whether the share of off-grid, behind-the-meter, and solar-plus-storage PPAs in data center projects continues to rise.