Copper Miners Are Too Cheap

Copper is at a record and the developers who will mine it trade as if it were worth a third less. Marimaca, Hot Chili and the takeover math that closes the gap.

Copper fell almost 4% today. Silver fell more than 5%, gold another 1.75%, the whole hard asset complex red on one screen on one afternoon. I’m posting this into that red on purpose, because a down day is the best way to make the argument I’m about to make.

Copper is still just above $6. The companies that will dig the next decade of it out of the ground are priced as if it were worth a third less. Obviously, the easy trade on a red day is to call the gap a discount and buy the dip. But that ‘easy trade’ is misleading, and I would rather take the obvious objection apart myself than have a sharper subscriber do it for me.

A copper mine does not sell its metal at today’s price, or at any single afternoon’s quote. It sells fifteen years of it, at whatever copper averages over that life, so the number that sets the value of the asset is usually never spot.

It’s the long run price, and the long run price the whole industry uses is $4.25. Against spot, even after today’s fall, that is still ⅓ cheaper, this is not asymmetric. That is the bear case, it’s correct, and it means the thesis was never a discount to spot.

So, is $4.25 the right long run price?

I think it’s too low, and everything that follows is why.

Let me explain what a ‘long price’ is, since it’s not that well known. The long run price is the all in cost of the next tonne the world has to build to balance the market.

Below that price the marginal mine does not get sanctioned, the supply does not arrive, and the price has to rise until it does.

If we look at what it takes to build a tonne of new capacity today, we see that the capital cost of new copper supply has climbed hard, from the low thousands of dollars a tonne for the simplest leach project to the mid thirties for a sulphide build in a difficult place, grades across the major districts are falling, and a new mine takes about seventeen years from first discovery to first metal.

BHP reckons the world needs on the order of ten million tonnes a year of new mined copper this decade, a figure that already counts depletion and grade decline, and the visible pipeline does not come close to filling it.

Grasberg’s mud flood pulled the better part of six hundred thousand tonnes out of the next two years of supply last September, and the mine will not be whole until 2027.

Replacing what the depleting majors stop making takes a scale of newbuild the market doesn’t understand, on the order of the 80 new mines UNCTAD claims the world needs by 2030, and the marginal tonne in that pipeline does not get built at $4.25.

That’s the hole I see in the consensus.

This is close to Rick Rule’s copper case, and he has earned the right to make it. He frames it as 30 years of underinvestment the industry cannot unwind in five, with the ten largest copper companies needing to spend on the order of $250 billion simply to hold production flat, and his line is that five years out we will be rationing copper by price.

He will also tell you, in the same breath, that he has no idea when it pays, that the move is measured in years and not weeks.

Here’s the test. If the long run price is right at $4.25, the majors should be sanctioning new builds at $4.25 economics. Spoiler, they aren’t. They’re paying up for finished assets instead.

The same issue runs through the whole ‘hard asset’ complex.

The metal gets crowded and the equity gets ignored, in gold, in uranium, in silver, in copper.

This phenomenon is something I have been following for over a year now in TSCS, specifically, I’ve done analysis on gold, silver, and uranium producers.

For interest sake: The gold miners pricing the metal a third under spot, name by name. The uranium names pricing the term market and shrugging off the spot panic. The silver basket hollowed out by acquisition and curated down to the few worth holding. None of those is a back issue.

Gold has moved hard since I wrote that piece, the miners have flushed, and a live position has to move with the facts, so we update each of them as the rotation develops rather than letting them set.

So how does one read the long run price off a screen? You find the flat copper price at which a project’s after tax study NPV equals what the market pays for the whole company today, and you read it off name by name, the way I did the gold miners a fortnight ago.

But there’s one difference with copper. A producing gold miner has no mine left to build, so its number is close to a pure view on the metal. A copper developer still has to fund and build the thing, so part of its discount is the market pricing the capex, the dilution and the years. That difference tells you which names are a clean read on the price and which are carrying a financing story inside the number, and the two are not bought for the same reason.

Marimaca is an interesting one. Copper only oxide, no gold to flatter it, a build that costs 587 million against a study already through Chilean environmental approval, so the execution risk is about as low as a developer gets. The study puts the project at 709 million dollars at 4.30 copper and 347 million at 3.50. The market pays 594 million for the company. Run the company’s own two points against its own price and the market is buying Marimaca right around four dollars.

If the deck is right, Marimaca is fair, and you earn your cost of capital and nothing on top.

If the deck rises, Marimaca re-rates one for one with it, the cleanest claim there is on the long run price being wrong. You are not buying a mispriced asset. You are buying the deck, in which this specific name seems to have the least going against it.

Hot Chili is also compelling, and it’s the other kind of name. Its study puts the project at 1.2 billion at 4.30 copper and 2.2 billion at 5.30, roughly 100 million of value for every ten cents of copper, and spot today sits above the top of that published range. The market pays 296 million for the company. At the study’s own base case the project is worth four times the entire company, and that is the figure I’d put weight on.

But be honest about why the gap is that wide, because it is not all copper. Building Costa Fuego costs 1.27 billion, more than four times the market cap, so a real slice of that discount is the dilution it takes to get there. The study still credits gold at 2,280 while gold trades north of four thousand, and it ignores the water business built to fund the mine, so there are offsets the other way. But the core fact stands. This is a name where the discount is part copper view and part financing problem, and a name like that doesn’t re-rate its way out. It finds its value through a buyer. Which is the next part of the story.

There’s a reason these are the only names left to run the exercise on. The independent copper developer is a dying category. BHP and Lundin took Filo out at the start of last year, folded it into the Vicuña ground, and crossed one more name off a list that was already short.

Every good asset that cannot fund itself is now an acquisition waiting to happen, because the world needs the tonnes and cannot build them fast enough to wait.

Watch what just happened to Arizona Sonoran. The market spent years pricing that company as if copper were $3. In March, Hudbay agreed to buy the whole thing, at a 30% premium to where it had been trading, an enterprise value of about 1.3 billion against a study worth 2.3 billion at 4.25 copper. 6/10 of NAV at a conservative deck is, on its own, an ordinary takeout multiple, because an acquirer discounts for exactly the capex and risk it’s taking on.

That’s the point. Do you get it yet?

A major does not face the dilution that caps the standalone equity. It funds the build off its own balance sheet, so it can pay a price the public market never would and still call it a bargain.

So there are two separate ways to get paid, because the whole thesis blurs if you don’t. The first is the deck. If the long run price rises, every one of these NAVs lifts at once, the clean names like Marimaca most directly, and it’s slow.

The second is the takeout. It does not need anything to move at all. It needs an acquirer with a fifteen year hole in its own pipeline and a balance sheet to make a financing problem disappear, both of which the majors have, and it arrives as a premium to a price the minority shareholder thought was already too low.

The discount that reads as a financing risk to a minority holder reads as a buying opportunity to the one buyer who doesn’t share the risk.

That’s how the repricing in copper developers actually shows up. Not as a re-rate the market grants you, but as a takeover the seller’s own shareholders will, in hindsight, wish they had not needed.

where I’m wrong

I’ve already conceded the big one at the start. The gap to spot is not a discount, and anyone who buys this thinking spot makes these names cheap has misread copper.

So the bear case is not about spot.

Copper pulled back nearly 4% today and is lower on the month, on Rio Tinto restarting Oyu Tolgoi and a Fed signalling it’s probably not done.

The dollar is at a 13 month high and leaning on every commodity priced in it. And the one buyer that sets the copper price is China, more than half of global copper demand on its own, the same economy whose property and industrial demand can wobble without notice.

If China sneezes, spot will tank, and a soft spot price drags this down with it in the market’s mind whatever the cost curve says. Worse, it does it at the exact moment these companies are trying to fund.

If copper softens while the developers are raising, the dilution guts the existing equity, and Hot Chili at a third of asset value turns out to have been correctly marked for the raise rather than cheap on copper.

The failure mode is not that I have the cost math wrong. It’s that the market prices the long run off the spot tape for long enough to kill the equity before the cost curve has a chance to be right.

I take that seriously, and here’s why I still lean, as a lean and not a certainty.

A Chinese soft patch won’t build the mines the world is short, so it doesn’t close the structural gap, it only delays the price that has to rise to get the tonnes built.

And the financing risk the bear is right to fear is the very thing that produces the takeouts. The assets that cannot fund themselves get bought by the ones that can, and they do not wait for a friendly tape. The Hudbay deal did not.

So the kill switch is clean. If copper breaks toward four and holds there, on a firm dollar and soft Chinese data, and the developers start printing financings into a weak tape, then the deck was right at four and a quarter all along, the cost curve I’m pointing at did not bind in time, and I was early at best.

What I’m watching are those same two dials, because between them they decide whether the deck gets to rise or the equity gets dragged down to meet a four dollar world first.

One thing to be straight about, because I would rather lose a subscriber than mis-sell one.

This does not pay on your calendar. It can sit, and it can drawdown first, gold and the gold equities especially, because in a liquidity squeeze the margin clerk sells whatever still has a bid.

Rule puts the warrant in years. I have high conviction on the direction and none at all on the date, and anyone selling you the date is selling you something. I’m a long term investor, I don’t care about your perfect entry point.

And if you take one thing from the whole rotation, take this.

Three of these four legs lean on the same cycle, copper and uranium and the grid and the build out, all of it ultimately a wager on global industrial demand with China at the centre of it, which makes them far more correlated than they look sitting in a basket dressed up as diversification.

Gold is the one that doesn’t.

It runs on debasement and on what central banks do with their reserves, which is exactly why it is the true diversifier in the set, the leg that tends to pay when the other three are busy not paying. The miners are cheap across the board.

The reason to own the metal under each of them is not the same reason every time, and forgetting that is how people end up holding four versions of one bet and calling it a portfolio.

The subscription is cheap on purpose. It buys you a seat before the deck reprices and the work to build the position while it is still unloved. It doesn’t buy you a discount to spot, because there is not one. It buys you a leveraged claim on the long run copper price being too low, in the names where that claim is cleanest, and the patience to hold it until the cost curve makes the case the tape will not. Judge the work, not the pitch.