We Bet Against Lithium, Then Bought Two Lithium ETFs

On July 3 we published a thesis: sodium-ion batteries are the next leg of the energy build-out. Sodium is roughly a thousand times more abundant than lithium and costs $100–500 a tonne against lithium’s $6,000–83,000. It runs at −40°C and it doesn’t burn. The play, we said, was picks and shovels — in a gold rush you don’t dig for the gold, you sell the shovels.

The basket: GM (industrializing the cells), FCX (copper, the input nothing electrifies without), and two thematic ETFs, BATT and LIT.

Three weeks later the market has answered, and the answer is uncomfortable in a specific and instructive way.

The scoreboard

Name	Price	vs 50d	52w position	Read
GM	79.52	+0.9%	53%	Working
FCX	62.56	−1.8%	55%	Working
BATT	14.43	−9.9%	17%	Broken
LIT	69.08	−13.7%	10%	Broken

Two are constructive. Two are near 52-week lows, about 10% under their 50-day.

And the two that broke are both lithium ETFs.

The thesis called its own casualty

Read that back against what we wrote. The argument was that sodium displaces lithium on cost and abundance. Then we put two lithium-weighted ETFs in the same basket as the trade expressing it.

We were long the shovel and long the gold at the same time, and we called it one idea.

The market did not disagree with the thesis. It resolved it. If sodium-ion is genuinely the cheaper chemistry, then a basket of lithium miners and processors is the wrong wrapper for that view — not because the electrification story is wrong, but because those companies sit on the side of the trade the story is arguing against. The half of the basket that broke is the half our own analysis predicted would break.

GM and FCX are a different exposure entirely. GM industrializes whichever cell chemistry wins. Copper is required by all of them. Those are shovels. BATT and LIT are a bet on one specific ore staying expensive.

What this does not mean

It does not mean lithium is dead or that we called anything perfectly. Three weeks is noise, two names is not a sample, and a supply glut has an expiry date — lithium ETFs bottoming and re-rating in 2027 would surprise nobody, us included.

What it means is narrower and more useful: a thesis and the instruments expressing it are two separate decisions, and we conflated them. Getting the macro call right and the wrapper wrong still loses money.

So the revision is not “sell lithium.” It is: the sodium thesis stays, and it should be expressed through companies that are chemistry-agnostic (GM) or chemistry-independent (FCX). For BATT and LIT there is no level worth naming yet — when a Fibonacci retracement zone sits above the current price, as it does on both, the retracement already failed and every “support” below is a step on the way down. The honest marker is LIT reclaiming its 50-day at 80.03. Not before.

And the part we changed in the machine

This same review surfaced a defect in CHESKO 2.0, our nine-agent desk, and it is worth publishing because it is the kind of gap that survives for months if nobody looks.

The desk is instructed, in its own prompt, to “think like a professional accumulating, not a market-buyer” — to stage entries into tranches at Fibonacci and support confluence rather than buying the first tick. And it did. Running it live on MTDR it returned a three-part ladder: 30% at 54.08, 35% at 52.58, and the heaviest 35% at 51.67, the golden-pocket floor.

Then execution ignored all three and bought 100% at market.

The ladder was computed, published to the dashboard, shown to followers — and never used. The desk was told not to be a market-buyer and was, on every single trade, a market-buyer.

That is now fixed, and split by horizon rather than given one answer:

• Day and swing entries stay at market. Over hours, missing the move costs more than paying the spread.
• Position and macro entries stage the ladder — an anchor tranche at market, the rest resting as GTC limit orders at the computed levels.

The trade-off is real and we are not hiding it: those resting tranches may never fill. If a name runs from here without retracing, the book ends up with the anchor instead of the full position. Over a months-long horizon that is a price worth paying; over a two-day trade it is not. Hence the split.

For the names above it means something concrete: the desk will no longer chase FCX at 62.56. It will put its weight where the golden pocket and the 20-day agree, near 61, and wait — and accept that sometimes the wait costs it the trade.

Why we publish the revision

Anyone can publish a thesis. The test is what gets published three weeks later when half of it is underwater, and whether the mistake gets named precisely enough to be useful — not “we were early,” but “we picked the wrong instrument for our own argument.”

Every position CHESKO takes, including the ones that go against us, is on a public page updated daily: balladmarkets.com/chesko. Open trades, closed trades, and the drawdown next to the return.

---

Levels calculated from daily closes through July 21, 2026. Ballad Markets is a market analysis and information service, not personalized financial advice. Levels are analytical reference points, not recommendations, and every one of them can fail. You are responsible for your own risk. CHESKO 2.0 runs on an Alpaca paper account: real orders, real fills, no real money.