Most lithium coverage focuses on mining. New deposits, new mines, new entrants. Mining is the visible, geopolitically charged, photogenic part of the chain. It’s also not where the bottleneck is.

The bottleneck is refining.

The concentration problem

Roughly 70% of the world’s refined lithium (both lithium carbonate and lithium hydroxide) is processed in China. This is true even for ore mined in Australia, the Americas, or Africa. Ore is shipped to China, refined there, and refined product is shipped onward to cell-makers globally.

This is a single-point-of-failure for everything downstream. The Q1 2026 price action made the asymmetry visible: battery-grade lithium carbonate prices nearly doubled from December to late January, driven primarily by a structural deficit in spodumene combined with the suspension of CATL’s Jianxiawo mine and Zimbabwe’s late-February export halt. The spread between spodumene feedstock and refined carbonate output is now the cleanest single signal on chain-tightness, and it sits squarely on the converter (refining) side, not the mining side.

Why refining is sticky

Three reasons concentration is hard to unwind:

  1. Capital intensity. A modern lithium-hydroxide refinery runs $500M–$1B+ in capex. Permitting, construction, and ramp typically take 3–5 years. There aren’t shortcuts. Lithium Americas’ Thacker Pass processing facility has $1.3–1.6B of capex just for 2026 phase-1 work, financed in part by a $2.23B DOE ATVM loan at Treasury rate + 0% spread.
  2. Process complexity. Lithium chemistry varies by feedstock (spodumene vs brine vs clay vs sedimentary). Each requires different process trains. Chinese refiners spent 15+ years building this expertise; replicating it is multi-year.
  3. Skill base. Operating talent for these facilities is concentrated in a handful of teams globally. The labor flywheel takes time to build elsewhere.

Australia is illustrative. The country has been pushing hard since the late 2010s to convert from raw spodumene exports to value-added domestic refining, and as of 2026 still faces “technical commissioning hurdles” per the Inside Climate News reporting on the broader lithium buildout. Even with capital, geology, and political will aligned, building refining capacity at scale is a years-long execution problem.

What’s actually changing

The IRA, DPA invocations, and DOE loan guarantees are routing real capital to non-Chinese refining. The marginal dollar of new capacity is mostly going to North America, Europe, and Australia. Specific things to track:

  • DOE Loan Programs Office decisions on lithium-refining projects. Conditional commitments are leading indicators; final closings are lagging. The first $435M drawdown on the Lithium Americas / Thacker Pass loan in early 2026 is the largest concrete example of capital flowing into US-domiciled refining capacity.
  • DOE Office of Critical Minerals funding rounds, including a $69M opportunity announced in April 2026 for next-generation processing technologies (DLE, alternative spodumene conversion, brine-to-hydroxide pathways).
  • IRA Section 30D FEOC rules and how they’re applied to specific offtake structures.
  • State-level permitting friction. Nevada, North Carolina, Tennessee, and Quebec are the active permitting battlegrounds.
  • Australian and Canadian refining buildout. Not US, but FEOC-clean and IRA-eligible.

What positioning looks like

The investable surface here is narrower than mining. Names with credible non-Chinese refining capacity in the planning or build phase number in the dozens, not hundreds. Most are early-stage. A few are pure-play. The rest are subsidiaries of larger mining or chemicals businesses.

The asymmetry: any name that achieves operational refining capacity at scale, FEOC-clean, with IRA-eligible offtake, is structurally advantaged for a decade. Most names will not get there. The screening question is which ones will, and the indicators are capex commitment relative to total cost, DOE loan-program engagement, EPC partner quality, and regulatory-clean construction-progress evidence (not just press-release optimism).

That’s the next analysis.

refiningpolicysupply