The Battery Behind AI: Why BESS Supply Chains Are the New Data Center Risk

The supply chain runs through five distinct layers, each with its own geography, chokepoints, and strategic risk: raw critical minerals; midstream refining and processing; cell manufacturing; the system integration around the cell; and final deployment. Lithium iron phosphate (LFP) chemistry now dominates stationary storage — more than 90% of new grid-scale capacity — having displaced nickel-manganese-cobalt (NMC) on cost, cycle life, and thermal safety.

Layer 1: The Critical Minerals

Every lithium-ion battery depends on lithium, plus — depending on chemistry — graphite, phosphate, iron, nickel, manganese, and cobalt. Lithium is reasonably abundant geologically but concentrated in production: Australia (hard-rock spodumene), Chile and Argentina (brine), and China dominate mining, and 2025 prices remained depressed after the 2023–2024 crash, masking a structural tightening expected later this decade. The more acute vulnerability is graphite: China controls over 90% of anode-grade graphite processing, and both natural and synthetic graphite anodes route through Chinese supply almost universally. Phosphate and iron for LFP cathodes are abundant, but battery-grade purification is again concentrated in China.

Layer 2: The Midstream Chokepoint

Midstream refining is where China's dominance is most complete and most strategically consequential. China refines roughly 65% of the world's lithium, 68–75% of its cobalt, and the overwhelming majority of battery-grade graphite, manganese, and phosphate — regardless of where the raw ore is mined. This is the true chokepoint of the entire chain: a country can mine lithium in Australia or cobalt in the DRC, but the material still flows through Chinese processors before it becomes a battery. Building non-Chinese refining capacity is slower and more capital-intensive than opening new mines, which is why the midstream, not the mine, is the binding constraint on Western supply diversification.

Layer 3: Cell Manufacturing

Cell manufacturing is dominated by a small number of Asian giants. CATL alone holds roughly 37% of the global battery market, and with BYD the two Chinese leaders account for more than half of all cells produced. Korea's LG Energy Solution, Samsung SDI, and SK On, plus Japan's Panasonic, make up most of the non-Chinese balance. CATL's LFP cells are the de facto global standard for stationary storage, and its technology licensing — including the arrangement underpinning capacity at U.S. facilities — extends its reach even into ostensibly domestic production. U.S. cell manufacturing is scaling rapidly under IRA incentives but still represents a small fraction of global output, and almost every U.S. line depends on Chinese equipment, materials, or licensed technology.

Layer 4: The System Around the Cell

A deployed BESS is far more than its cells. Around them sit the battery management system (BMS), the power conversion system (PCS/inverter), thermal management and liquid cooling, fire detection and suppression, the enclosure or container, and the energy management software that dispatches the asset. In a representative 100 MWh system at about $175/kWh, cells account for roughly half the cost, with the PCS near 10%, containers and thermal/fire systems near 18%, AC/DC balance of system and grid interface near 10%, and engineering and EPC near 12%. The PCS layer carries its own geopolitical exposure:

suppliers dominate globally, and Chinese-origin units make up the majority of the PCS on California's BESS allow list — the same security concern now driving proposed U.S. inverter restrictions.

Layer 5: Deployment Economics

At the deployment layer, the economics have become compelling: falling cell prices, the standalone-storage investment tax credit preserved under the 2025 OBBBA, and surging demand from data centers, renewable firming, and grid resilience. Texas (ERCOT) and California (CAISO) lead U.S. deployment, with ERCOT's energy-only market and price volatility making merchant storage especially attractive. But developers face a tariff and trade-policy environment in flux — Section 301 tariffs on Chinese batteries, FEOC restrictions that disqualify Chinese-linked content from certain credits, and the prospect of PCS and inverter import limits — all of which reward early supply-chain diligence.

The Bottlenecks That Matter

The binding constraints are concentrated in the midstream and in policy, not in geological scarcity. China's refining dominance is the single greatest structural risk. Graphite anode processing is a near-monopoly. FEOC and UFLPA-style compliance make provenance documentation a gating requirement, not an afterthought. Transformer and grid-interconnection lead times — shared with every other clean-energy sector — routinely delay BESS projects by years. And the PCS layer inherits the same Chineseinverter security scrutiny reshaping the solar market. None of these is a simple hardware shortage; each is a concentration or policy risk that rewards diversification planned years ahead.

What Developers Should Do Now

Practical steps follow directly from the layered risk map. Document cell and material provenance early to satisfy FEOC and taxcredit requirements, and qualify Korean or Japanese cells, or IRA-compliant U.S. production, where credit eligibility matters. Treat the PCS as a strategic sourcing decision, evaluating Tesla, SMA, Parker/Dynapower, and Power Electronics alongside Chinese incumbents. Order transformers and secure interconnection positions 24–36 months ahead of commissioning. And build tariff and trade-policy escalation into project models as an ongoing condition rather than a one-time shock — the regulatory environment for storage will keep moving faster than the hardware.

Sources