Notes
Battery Technology & Grid Storage
On this page
The battery — long the bottleneck of the energy transition — has become its linchpin. Volume-weighted lithium-ion pack prices have fallen roughly 93% in real terms since 2010, riding an experience curve of about 18% cost reduction per doubling of cumulative production, and that single curve now determines whether EVs reach true mass market, whether wind and solar can behave like baseload, and whether the AI buildout can be powered without a reliability crisis.
The steepest useful curve in energy
The trajectory is concrete: roughly $1,400–1,500 per kWh in 2010, $159 in 2020, $108 in 2025 — with EV cells averaging $79/kWh and Chinese packs at $84. The curve survived a lithium price spike and correction in 2022–23 and is projected to keep sloping toward $40/kWh by 2035, with annual demand passing 2 TWh. The consequences arrived on schedule: many EV segments crossed price parity with combustion vehicles (18 million EVs sold globally in 2025), and solar-plus-storage PPAs in the best regions signed below $40/MWh all-in — cheaper than new gas almost everywhere.
Chemistry diversifies, manufacturing gets attacked
Lithium-ion is no longer one thing. LFP — cheaper, safer, 6,000+ cycles in stationary use — now exceeds 55% of global EV batteries and roughly 90% of grid deployments. Sodium-ion moved from lab curiosity to first commercial fleets: CATL's second-generation cells target 200–220 Wh/kg using abundant salt-derived materials and retain about 90% capacity at −40°C, a structural hedge against lithium geopolitics. Solid-state transitioned from hype to pilot production, with China's formal standard taking effect in 2026 and 300–500+ Wh/kg cells targeted for premium vehicles in 2027–28 and mass market in the early 2030s.
The most consequential near-term advance may be a process, not a chemistry: Tesla's solvent-free dry electrode production for both anode and cathode of its 4680 cells, now in actual production vehicles. By eliminating toxic solvents and football-field drying ovens, it points toward electrode cost reductions approaching 50% and pack-level savings of 20–30% — an industrial play that multiplies the advantages of whatever chemistry is coated onto the foil.
From peaker replacement to grid backbone
Grid-scale storage stopped being a demonstration: 49.4 GW / 136.5 GWh came online in just the first nine months of 2025. The services matured with the scale — batteries at Moss Landing and across ERCOT respond to frequency events in under 50 milliseconds while simultaneously arbitraging energy, and grid operators now routinely specify grid-forming inverters that supply the synthetic inertia retiring thermal plants once provided. In Texas, storage clears the evening real-time market more often than gas. For the durations short-duration LFP can't economically serve, iron-air and flow batteries are winning 10–100+ hour contracts — a 2026 Inner Mongolia project pairs 200 MW of wind with 1.6 GWh of 100-hour storage to deliver firm power around the clock.
The essays' through-line: the technology is no longer the limiter; the institutions are. Interconnection queues exceeding three years, permitting, and transmission buildout are now the binding constraints — the same ones throttling EV charging infrastructure and data center interconnects. The battery revolution is no longer coming. The question is how fast societies can build the factories, mines, recycling systems, and grid infrastructure to keep up with the demand they created.
Related concepts
Appears in
Details
- Section:
- Notes
- Updated:
- 2026-06-12