This is the founding thesis for the energy storage and lithium vertical. Clean Power Press now covers five verticals; each has its own thesis frame. This piece covers storage and lithium specifically.

EV adoption was the most-covered demand story in lithium through 2024 and 2025. It was also the most overweighted in headlines and the most misleading as a positioning frame. The “EVs are slowing, lithium is oversupplied” narrative that drove the price collapse from late-2022 highs has fractured, and the data driving the reversal didn’t come from the EV cycle.

What the data is saying now

Battery-grade lithium carbonate spot prices ran from approximately US$13,433/t in early December 2025 to US$26,278/t by late January 2026, a 95% surge in roughly two months (S&P Global / Investing News). Both Morgan Stanley and UBS now publish 2026 deficit forecasts: Morgan Stanley at ~80,000 tons LCE deficit, UBS at ~22,000 tons. Both reverse the 2025 surplus.

The supply-side cuts driving the move:

  • CATL’s Jianxiawo mine in China remains suspended, removing meaningful marginal supply.
  • Zimbabwe announced a suspension of lithium-concentrate exports in late February 2026: a non-Chinese, non-Australian source losing spot availability in the same quarter as the price move.
  • Spodumene structural deficit. Converter overcapacity in China is now larger than upstream feed availability, tightening the spod-to-carbonate spread.

Where incremental demand actually comes from

Energy-storage demand grew approximately 71% in 2025 with analyst forecasts adding another ~55% in 2026 (BMI). Concrete data point: ERCOT entered 2026 with 13.9 GW of operational grid-scale battery storage and Modo Energy projects 40–55 GW operational by 2029. Texas-developer Greenflash Infrastructure has separately secured >10 GWh of BESS capacity earmarked for data-center power.

The structural demand vectors over a 10-year horizon, ranked roughly by marginal-effect-on-lithium-balance:

  1. Grid-scale storage. Largest single non-EV vector. ERCOT and CAISO are the leading edge; MISO and PJM are 2–3 years behind ERCOT’s 2022 inflection.
  2. AI-data-center backup. Per-rack lithium use is small; aggregate across the AI capex buildout is meaningful and not yet visible in most demand models.
  3. Robotics / humanoids. Bullish-case curves not yet in mainstream models.
  4. Behind-the-meter solar. Residential + small-commercial storage attachment rates climbing globally.
  5. EVs. Still real. No longer the only story.

Why supply is the more interesting question

US lithium projects average 7–15 years discovery-to-production. Thacker Pass (the canonical case) is now ~$1.3–1.6B in capex for 2026 alone with mechanical completion targeted for late 2027 (Lithium Americas press release). The capital cycle is also lagging: lithium prices crashed in 2023–2024, deferring marginal capex; the cycle that shows up in production lags 5–7 years.

Refining concentration is the stickier problem. ~70% of refined lithium routes through China today. The IRA, DPA invocations, and DOE loan guarantees are routing real capital toward non-Chinese refining (Lithium Americas has already drawn the first $435M tranche on a $2.23B DOE ATVM loan at Treasury rate + 0% spread), but the existing concentration takes time to unwind.

Positioning implications

  • US-domiciled miners and refiners with IRA-eligible offtake have a structural advantage that compounds over time.
  • Pure-play lithium ETFs (LIT, BATT) trade as a proxy for the whole story; blunt instrument.
  • Watch the marginal capex dollar, not the average inventory. Marginal capacity is heading to NA, EU, Australia.
  • Refining capacity, not just mining, is the chokepoint.

Risks to the thesis

  • Sodium-ion or solid-state batteries take grid-scale share faster than expected. Medium probability, large impact.
  • China retains export-policy levers and uses them strategically. Live risk, ongoing.
  • US permitting reform never lands meaningfully. Status quo keeps the chain bottlenecked.
  • Storage demand grows but lithium captures less of it than expected. Flow batteries, gravity, or thermal alternatives find their niches.

The frame: track marginal, watch policy, ignore most price noise.

thesisdemand