China’s non-conventional energy-storage capacity rose from under 4 GW in 2021 to approximately 155 GW in the first quarter of 2026. All ten of the world’s top battery-cell suppliers for energy-storage systems in that quarter were Chinese, together capturing 90% of the global market, according to Benchmark Mineral Intelligence. The U.S. ended 2025 with 57 GW of storage capacity; Wood Mackenzie projects 200 GW in five years, a pace that presupposes continued access to affordable components. The One Big Beautiful Bill Act bars storage projects using Chinese components from federal tax credits, and Trump tariffs have eroded the cost advantage that made Chinese batteries the default choice. The U.S. gap with China is 98 GW at the latest cited figures, and it is likely to widen before it narrows.

The structural asymmetry is not in the cell — it is in the processing. OBBBA and tariffs block Chinese cell imports at the final-assembly stage, but the binding constraint runs through the upstream processing of lithium, cobalt, and graphite. China holds a concentrated position in that processing capacity, and since 2023 Beijing has tightened export limits on those materials and on advanced battery technologies. Even if South Korean or Japanese suppliers — Samsung SDI, LG Energy Solution, Panasonic — scale their cell production to serve the U.S. market, those cells will still depend on Chinese-processed inputs. OBBBA closes the front door; the processing dependence is a side entrance no tariff can seal without disrupting the global battery supply chain. In game-theoretic terms, the stable subgame-perfect outcome is observed in the current state: the U.S. maintains high tariffs, and China cooperates at current export-control levels. China’s best response to further U.S. tariffs is not to escalate comprehensively, because a full cutoff would damage CATL’s revenue trajectory (storage targeted at 50% of global revenue by 2030, up from 15% in 2025) and would raise China’s own input costs. But that stability is conditional. If Korean and Japanese suppliers cannot reach competitive scale within three to five years, OBBBA becomes a costly bluff, and the U.S. move is exposed as a policy commitment without a production base.

Three reinforcing loops have operated unimpeded for five years. The first loop couples renewable installation to storage demand: solar and wind now account for 22% of China’s electricity, and the intermittency drives the storage buildout. The <4 GW to 155 GW expansion is the structural expression of this loop. Beijing’s 300 GW by 2030 target, announced under the 15th Five-Year Plan, is consistent with the same trajectory. The second loop is the scale-learning dynamic that produced the 90% market concentration: production volume drives down LFP unit cost, which drives up global price competitiveness, which drives demand, which drives further production volume and manufacturing scale. This is the dominant loop at all current timescales, and it is the reason that, in the words of Iola Hughes at Benchmark Mineral Intelligence, “no matter which market you’re in the world, the obvious choice would be to go for these batteries which the Chinese players had mastered.” The third loop adds a parallel demand driver from AI grid strain: Beijing’s mandate requiring all new data centers to derive at least 80% of their power from renewables creates a policy-anchored pull for storage. Even if solar and wind growth slows, AI-driven grid strain sustains storage demand through this loop. The U.S. policy damping loop — tariffs and OBBBA raising the effective price of Chinese cells in the U.S. market — operates only at the U.S. market boundary. It is structurally weaker than the three reinforcing loops because it is policy-activated rather than market-driven, because the U.S. has no domestic supply alternative at scale, and because the Tesla and Ford licensing pathways allow Chinese technology to enter the U.S. even under tariff constraints.

The Tesla and Ford dependencies are a partial-accommodation equilibrium that hollows out the tariff architecture. Tesla produces Megapack systems at its Shanghai factory using CATL cells for markets outside the U.S.; Tesla is the leading U.S. storage-system seller. Ford licenses CATL technology to produce storage products in the United States. The licensing pathway is a structural end-run around OBBBA: Chinese technology enters the U.S. under domestic-content rules that bar direct Chinese imports from tax-credit eligibility, sustaining CATL revenue while giving Ford and Tesla cost-effective supply. The Shanghai Megapack line is an irreversible investment; Tesla rationally continues using Chinese components even as tariffs rise — a firm-optimal decision that is nationally suboptimal. Whether this model scales is an open question: if other U.S. firms adopt similar licensing arrangements, the tariff-and-tax-credit architecture is hollowed out; if not, Ford and Tesla retain a privileged position that smaller competitors cannot match.

A fourth loop is latent within the Chinese system, and its resolution may matter more than any U.S. tariff. Robin Zeng, founder of CATL, warned in September 2025 of “vicious price competition” among Chinese suppliers, saying hard-pressed companies were cutting corners on quality. This signals a potential internal-constraint loop: price competition intensity drives quality compromise risk, which erodes buyer confidence, which dampens demand. The loop’s polarity is unresolved by the available evidence. If manufacturing scale tends to intensify price competition, the loop is balancing — a structural brake on Chinese dominance that operates from within, rather than as an external U.S. policy constraint. If scale tends to moderate price competition, the loop is reinforcing — a race to the bottom. No documented market-trust erosion event or field-failure data exists in the source; the loop is structurally present but latent. Its resolution will determine whether Chinese dominance cracks from within faster than U.S. policy can act from without.

Stakeholder salience follows the structural asymmetry. Beijing is definitive by all three Mitchell-Agle-Wood dimensions: it holds power through supply-chain control and export-limit authority, legitimacy as a sovereign policy actor with stated national energy targets, and urgency driven by the 2030 non-fossil target and AI-induced grid strain. Chinese battery suppliers are contested between dominant and definitive: the disagreement is on urgency, with one reading seeing non-U.S. market insulation as suppressing it and the other seeing CATL’s 2030 revenue target and the price-competition pressure as binding it high. The Trump administration is similarly contested between definitive and dominant: the disagreement is on whether the 2026 policy cycle and its immediate effect on project financing constitute high urgency, or whether the policy is reactive relative to China’s 2030 timeline and therefore moderate. Tesla and Ford are each contested between dangerous and dependent: the substantive disagreement is on power — whether market position or CATL dependency is the structurally binding dimension. Every other party is classified without ambiguity: U.S. storage project developers, grid operators, raw-material processors, and battery manufacturing workers are dependent; coal and gas generators and data-provider firms are discretionary; European battery manufacturers, developing nations, and U.S. electricity consumers are dormant.

The strategic recommendations that follow from the structural analysis are not symmetrical. For U.S. policymakers, the priority is to fund upstream processing capacity outside China through a coalition with South Korea, Japan, and the EU. This moves the game from bilateral rivalry to multi-player bargaining and addresses the zero-sum locus — processing — rather than the visible but secondary locus of final assembly. Direct subsidies for domestic LFP manufacturing, not just conditional tax credits, would create sunk-cost commitment that changes China’s expectation of U.S. durability. Maintaining tariffs with a publicly announced gradual phase-down schedule would alter the shadow-of-the-future discount factor, signaling that decoupling is temporary and reducing China’s incentive to retaliate aggressively while forcing immediate supply diversification. For Chinese policymakers, the optimal move is to expand the structured licensing model that Ford-CATL has already demonstrated: it preserves Chinese revenue, reduces U.S. dependence concerns, and undermines the coalition incentive by giving individual U.S. actors a better bilateral deal than collective action provides. For commercial actors, the dominant private-sector equilibrium given policy uncertainty is compliance-plus-hedging: source OBBBA-compliant supply for U.S. projects while maintaining Chinese relationships for non-U.S. markets, capturing the first-mover premium in non-Chinese processing if U.S. policy holds and hedging against reversal by preserving Chinese supply lines.

Four open questions will determine which trajectory the system follows. First, can Korean and Japanese suppliers reach competitive non-Chinese LFP cell capacity within three to five years? Without it, OBBBA is a costly bluff. The source package provides no capacity commitments. Second, does the OBBBA regime catalyze a U.S. domestic cell-manufacturing buildout, or does it produce slower deployment at higher cost without a corresponding industry? Wood Mackenzie’s 200 GW projection assumes continued deployment, but that pace is not assured by current conditions. Third, does the Ford-CATL licensing model scale? If other U.S. firms adopt similar arrangements, the tariff architecture is hollowed out; if not, Ford and Tesla retain an uncompetitive advantage. Fourth, does Zeng’s “vicious price competition” warning translate into quality failures that erode buyer confidence? The latent internal-constraint loop will determine whether Chinese dominance cracks from within faster than U.S. policy can act from without. None of these questions are answered by the available evidence. Each is structurally consequential for the trajectory of the global battery-storage market, the energy transition timelines of every importing country, and the viability of the U.S. attempt to decouple from a supply chain it depends on far deeper than the tariff line suggests.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Stakeholder Mapping
Charts the parties to a situation — their interests, power, and alignments.
Strategic Interaction (Game Theory)
Models a situation as a game — players, moves, payoffs, and likely equilibria.
Systems Dynamics (Structural)
Maps a system’s structure — stocks, flows, and the architecture that shapes its behavior.