Tearing apart the lithium restart mirage

Lithium.

The Guangzhou Futures Exchange just printed a five-month low. Every financial headline from here to Hong Kong is running the same take: CATL restarted the Jianxiawo mine, the glut is back, lithium is dead again. Short it.

I've seen this movie. I was sitting in a version of it in 2020 when copper did the same thing—everybody read the surface, nobody checked the loading docks. Here's what actually happened last month: one mine with 100,000 tonnes of annual capacity got a permit stamp on June 29, and the futures market priced it as though a geological miracle had materialized from the earth's crust. It hadn't.

The paper market saw one mine restart and priced in an entire supply thesis. The physical market—the converters, the cathode plants, the people actually buying lithium salts for battery-grade use—is telling a completely different story.

The surplus everyone keeps citing? It was consumed from the inside. Not by fresh production. By drawing down stockpiles that are now approaching empty. The market didn't rebalance. It just ran out of places to hide the deficit.

Lithium.

The Guangzhou Futures Exchange just printed a five-month low. Every financial headline from here to Hong Kong is running the same take: CATL restarted the Jianxiawo mine, the glut is back, lithium is dead again. Short it.

I've seen this movie. I was sitting in a version of it in 2020 when copper did the same thing—everybody read the surface, nobody checked the loading docks. Here's what actually happened last month: one mine with 100,000 tonnes of annual capacity got a permit stamp on June 29, and the futures market priced it as though a geological miracle had materialized from the earth's crust. It hadn't.

The paper market saw one mine restart and priced in an entire supply thesis. The physical market—the converters, the cathode plants, the people actually buying lithium salts for battery-grade use—is telling a completely different story.

The surplus everyone keeps citing? It was consumed from the inside. Not by fresh production. By drawing down stockpiles that are now approaching empty. The market didn't rebalance. It just ran out of places to hide the deficit.

Numbers. No story.

Lithium carbonate hit over $81,000 per tonne in late 2022. By mid-2025, it had cratered about 90% to below $10,000. That crash wasn't a correction. It was an extinction event for the mine investment pipeline.

Here's the mechanical sequence nobody has patience for.

Between 2022 and 2024, midstream converters in China stockpiled aggressively during the price spike. When prices collapsed, they stopped buying spot and burned through existing reserves instead. For eighteen months, consumption outpaced new production but nobody noticed because the stockpiles absorbed the difference. Benchmark calls it the "invisible deficit." By early 2026, those hidden inventories hit critical lows. Converters are now returning to the spot market—simultaneously.

Global lithium production reached roughly 1.5 million tonnes LCE in 2025. Demand growth from EVs and energy storage is running 30–40% year-over-year. Battery energy storage alone is forecast to account for 30% of global lithium demand by 2026—up from a rounding error five years ago. CATL just reported energy storage at 19.2% of its total revenue in H1 2026, up 87.5% from a year earlier. That's not a niche segment anymore. That's a second demand engine nobody budgeted for.

And the supply pipeline to match? Gutted. Albemarle shuttered its entire Kemerton hydroxide plant in February 2026. Arcadium announced plans to place Mt Cattlin into care and maintenance. MinRes shut Bald Hill. A greenfield lithium mine takes five to seven years from final investment decision to production. Projects needed for 2028–2030 required capital commitments during 2023–2025. At sub-$10,000 prices, those commitments didn't get made. The math is now locked in.

One mine restart—even one producing 100,000 tonnes annually—does not fill a structural hole that took three years of underinvestment to dig.

Here's what nobody in the futures pit is talking about: conversion capacity.

Mining spodumene ore is step one. Converting it into battery-grade lithium carbonate or hydroxide is where the system actually bottlenecks. You can pull rock out of the ground in Western Australia all day long. If you can't process it into a chemical that a cathode manufacturer will accept, you've got a very expensive pile of gravel.

Spot prices for spodumene concentrate have outpaced lithium hydroxide and carbonate over the past year. When the raw input rises faster than the finished product, it signals mine-level tightness feeding into a conversion system already under strain. Conversion margins of $13,000 to $16,000 per tonne signal a processing bottleneck, not a surplus.

And now consider who just arrived at the table. The GFEX opened lithium carbonate futures to overseas traders on July 3. Fresh speculative capital from people who've never toured a conversion plant in Jiangxi and wouldn't know lepidolite from limestone. They see the Jianxiawo headline, run the simple math—100,000 tonnes of new supply!—and hit the sell button. They don't ask whether that mine produces battery-grade material directly. It doesn't. It produces lepidolite concentrate that still needs to be converted. Into a conversion system that's already running hot.

I watched the same pattern in nickel in 2021. Indonesian mine supply flooded the headline count. Prices softened. Then everyone realized the bottleneck was in Class 1 nickel processing, not in laterite ore. What followed was—let's call it educational for the shorts.

The invisible deficit becomes visible in stages. You've seen the choreography before if you've watched any physical commodity squeeze in the last five years.

Stage one is already underway. Converters are returning to spot after eighteen months of running down reserves. Stage two: conversion margins keep widening because the bottleneck is processing, not ore. Whoever has battery-grade output under contract has pricing power. Everyone else is competing for scraps.

Stage three is where it gets violent. The GFEX contract is down 30% from its May high. Fresh overseas speculative shorts are piling in on the glut narrative—many of them brand new to this contract since July 3. If physical tightness forces a spot repricing, the way it did in Q1 when lithium carbonate pushed above $24,000, those shorts become fuel for the move. The mechanics are identical every time. Only the trigger changes.

Where does the capital go? Not into the broad lithium ETFs. They're diluted with explorers who won't produce for years and processors still running at a loss. The edge—if there is one—is in integrated producers with both mine output and conversion capacity. Companies that control the entire chain from ore to battery-grade chemical. They're the ones whose revenue rises on both the physical squeeze and the conversion margin blowout. Everyone else is selling spodumene into a bottleneck and hoping someone else can process it.

The futures market says lithium is going down. The converters say they can't get enough battery-grade material. When those two disagree, I've learned—sometimes expensively—to bet on whoever's actually handling the physical product. It takes longer than you want. It usually pays better than you expect.