Welcome back to CMJ,

20-second highlights:

European erbium is up more than 50% since June. No fibre network changed its build plan in eight weeks. What changed is a date, and the buying is what that position looks like in physical form.
Chile stopped taking its dividend from Codelco for the first time since 1976. That is a fiscal decision before it is a mining one, and it tells you what the state now thinks the copper cycle is worth.
The LME intervened on August 14 to contain the price of immediacy. Three days later the spread was about US$110 a tonne wider than when it intervened.
The loudest copper number of the week appeared on no exchange screen at all. It is a treatment charge, and it has fallen US$128 a tonne since January 2.
Korea bought copper security for US$1bn and disclosed no tonnage, no price and no term. The counterparty was not a mine.
If November 10 moves, several positions taken this week unwind in the same direction on the same day, across minerals that have nothing to do with each other. Nobody is describing that as a correlation, and this week it became one.

Illustration of South Korea's Cheong Wa Dae (the executive office and residence of the president of South Korea)

Three markets bought the same thing this week, and it was not metal

First, Erbium (a rare-earth metal used in telecoms).

Erbium goes into the doped fibre amplifiers that keep long-haul optical links from fading, it has close to nothing to do with permanent magnets, and for most of the past decade it has been one of the least interesting prices in the rare earth complex.

It has been reported that Erbium’s European price reference has risen by more than 50% since June, with Chinese prices up roughly 40%, as buyers accumulate ahead of the November expiry of the suspension covering parts of China's expanded export-control regime. Holmium and ytterbium have drawn the same precautionary buying.

Nothing happened to erbium demand in eight weeks. Fibre build cycles do not turn that fast, and no producer announced a disruption.

What changed was proximity to the deadline, and it seems to be enough to spike prices.

Brief recap: Erbium, holmium, thulium, europium and ytterbium are the five elements China added to its control list on October 9, 2025, and then suspended until November 10, 2026. The buying is a position on the expiry, taken in warehouse form because there is no other way to take it.

None of that is remarkable or news on its own. A firm facing a possible licensing regime in twelve weeks buys inventory now (for stockpilling), and every procurement desk in this sector has run that arithmetic at least once this year. This is the new norm.

For those that are not fully aware, the date itself is worth stating precisely, because a great deal now hangs off it. China published the controls on the five additional elements on October 9, 2025, with effect from November 8. In early November, following the Busan meeting, it suspended them until November 10, 2026, and the U.S. suspended the BIS Affiliates Rule for the same period. Neither instrument was withdrawn. Both were parked, with a shared expiry, and the parking space runs out on November 10, 2026.

A suspension with an expiry is a different object from a repeal. It leaves the rule drafted, the licensing apparatus built and the decision entirely discretionary, which is precisely the structure that makes a forward position possible. Buyers here are taking the other side of an option China owns and has not said whether it will exercise. That is a different exposure from a scarcity hedge, and it settles on a different kind of news.

Note that apparently there is a visit from China’s President Xi Jinping to the U.S. for three days, in which we cannot be certain, but we can infer with a good level of confidence that critical minerals are part of the agenda. It is scheduled to start on September 23rd.

And what makes it worth an edition is that two other markets did the same thing inside the same week, on two other dates, with no connection to erbium at the mineral level.

But Copper is the loud one.

On August 14 the August LME contract traded at a premium of as much as US$370 over September, the widest one-month spread since the 2021 squeeze, and the cash-to-three-month spread reached about US$434 a tonne. That prompted the exchange to introduce emergency measures to contain the rally in spot prices. By August 17 the cash-to-three-month spread was reported at US$543.50 a tonne, with cash near US$14,545 and three-month near US$14,134, and LME stocks down for a 42nd consecutive session to 204,975 t.

Read that sequence in order again. The exchange capped the price of immediacy on a Friday, and the price of immediacy was about 25% higher the following Monday.

It is worth being clear about what an intervention of that kind is:

Lending rules and backwardation limits do not add metal to a warehouse (of course).
But they constrain what a dominant holder of nearby positions may charge someone who has to roll or take delivery, which is a rule about conduct rather than about supply.

The exchange reached for the conduct lever because it is the only lever it has, and the spread widened past it anyway.

Underneath it sits yet another date.

Refined copper faces a phased U.S. duty from January 1, 2027, beginning at 15%, and metal has been migrating into U.S. warehouses against that possibility for eighteen months. The London pool is draining in part because the marginal tonne is sitting behind a tariff wall that has not been built yet.

There is a tell in how the copper position is held. Metal moved into U.S. warehouses ahead of a duty, so that inventory imported at zero would be worth the tariff-inclusive domestic price if the duty ever lands. The return on that trade comes from the announcement rather than from the copper, and it accrues only to metal already inside the perimeter when the decision is taken. That is why the pile kept growing through eighteen months in which nothing was decided.

So one market is buying an expiry, and another is buying a duty that may never arrive. Neither position is about the metal, and neither holder controls the outcome.

The third market closes the pattern, and it is the one nobody files under geopolitics. Once the three sit together, the way these positions come off becomes the more consequential question, because they come off on the same kind of news.