Argonne's BatPaC cost model (ANL/CSE-24/1, Table 36) puts contained lithium in an LFP pack at 0.09 kg/kWh. This is an automotive pack figure, not a measured 314 Ah ESS cell bill of materials, but it is the best publicly sourced primary number available. Convert to lithium carbonate equivalent via the stoichiometric ratio (x5.32) and the sensitivity coefficient is ~0.48 kg LCE per kWh.
InfoLink's August 5 battery-grade carbonate average fell CNY 4,000/MT WoW. Multiply through: 0.48 x CNY 4.00/kg = CNY 0.0019/Wh. That is 0.5% of InfoLink's CNY 0.360/Wh cell midpoint, and it sits inside the noise floor of both InfoLink's CNY 0.060/Wh assessment band and SMM's weekly 314 Ah index movement (CNY -0.0014/Wh over the same period). At this magnitude, the weekly lithium move falls below the resolution of the price assessments tracking cells.
Iron phosphate precursor, meanwhile, rose ~45% in H1 2026 to RMB 14,500-15,000/MT (SMM, July 13-16), narrowing whatever room lithium relief might create. The explanation for "cells didn't follow lithium" is the size of the impulse itself.
Calculation chain
- Li intensity: 0.09 kg Li/kWh (Argonne ANL/CSE-24/1, Table 36; LFP pack basis)
- Li to LCE: x5.32 (molecular weight ratio Li2CO3 / 2Li; stoichiometric)
- LCE intensity: ~0.48 kg LCE/kWh (derived)
- Weekly carbonate move: CNY -4,000/MT (InfoLink, Aug 5)
- Cell-cost impulse: 0.48 x 4.00 = CNY 0.0019/Wh
- Share of cell price: 0.0019 / 0.360 = ~0.5%
What this does not cover
- Argonne models 125 Ah automotive cells; 314 Ah ESS electrode loading differs, so true lithium intensity may vary
- Assumes immediate full pass-through: zero inventory buffer, contract lag, or hedging
- Single-week move only; cumulative declines (e.g., CNY -33,000/MT since early June) produce proportionally larger impulses warranting separate analysis
- Iron phosphate is a separate, rising cost input outside this calculation