The Copper Case and Its AI Demand Leg

2026-10-11 · The investor's Sunday · The copper case tested, Week ahead for a Nordic portfolio

Daily show9:53Download MP3About the show

How it is made. Aavistus Daily is written and voiced by our own systems, every morning, from data we gather ourselves. The voices are synthetic. Every fact is traced to a source on this page, and every dated call is settled in public on the calibration page. The presenters, Nora and Callum, are named characters, not people.

Aavistus is a new framework, still under development. We will make mistakes, and every dated call is settled here in public so you can see them; the aim is to keep calibrating until the record beats the average forecaster, not to sound right.

Testing The Finance Bureau's copper-as-strategic-metal case against the AI capex curve its demand depends on, with Joseph Wang and Lance Roberts arguing the capex curve breaks. Then the Nordic week ahead.

Aavistus briefs. Independent market intelligence — satellites, flows, filings. Register free to follow new briefs.

The copper case tested

Nick at The Finance Bureau this week argued that copper has stopped being a cyclical read on global growth and become a strategic metal defended by three structural buyers at once: artificial-intelligence data centers, grid build-out and electric vehicles.

Supply — the strongest leg

The supply side of the argument is hard to argue with. Nick points to declining ore grades, decades-long lead times for new mines, and operational disasters at major sites running at the same time. Three structural constraints and a near-term supply shock, compounding. That is a decade-long condition baked in regardless of what demand does.

Demand — turns on the AI capex curve

On the demand side Nick leans harder than the evidence supports. The medium-term demand case rests on the AI capex curve staying intact and on the grid-build curve not being cut back by national budget pressure. If either slows, the demand assumption underlying today's prices is a different picture.

Two voices arguing the AI capex curve breaks

Joseph Wang on Markets Weekly argued that soaring AI capital expenditures face an unsustainable revenue shortfall. His picture is one where hardware spend is dramatically out of proportion to the revenue currently generated by the services it supports.

The direction is fine. The timing is where Wang overshoots. The names carrying the AI concentration — Microsoft, Nvidia, Alphabet, Amazon, Meta — are producing real cash flow now, even where the return on their capex is not yet there. A gap between return-on-capital math and index-level pricing can persist for a long time if cash flow and backlog growth stay strong. These names have carried the broad US index over the past quarter while the equal-weighted index has lagged. That is not a tape rolling over on its own.

Nick on The Finance Bureau made a complementary point the same day. He argued headline AI revenue figures can be deeply misleading — that annualized run-rates computed from a single strong month, and the difference between gross and net revenue recognition through cloud distribution partners, both distort how large these businesses actually are. For a Nordic investor sizing US tech exposure, that is a reminder that the headline revenue figures driving current multiples are not all measured on the same basis.

Lance Roberts on the tape

Lance Roberts on Thoughtful Money with Adam Taggart came at the same question from a different angle. He argued that megacap tech has become extremely overbought and that he is trimming those positions back to benchmark weights, pointing to factor divergence between the cap-weighted and equal-weighted broad US indices. His case is a technical-tape case rather than a return-on-capital case.

We agree the equal-weighted-to-cap-weighted divergence is the mechanism of a slow index-level de-rating happening under the surface. We disagree with Roberts' near-term rotation into value, financials and utilities — the catalyst is not there inside the next quarter, with the Fed still hiking, the US long end near multi-decade highs, and financials more exposed to US-dollar funding-cost compression than to a growth rotation. Roberts and Wang disagree with each other on the mechanism; both can be right directionally and both overshoot on timing.

The read

Three arguments this week each identifying a different reason the AI capex curve might break, and none of them firing yet. The structural call that copper is supply-constrained through the decade stands on the supply numbers alone. The case for chasing the metal at current levels is harder, because strategic-reserve build-up and tariff front-running have already pushed inventory into government hands and that is partly what the price already shows.

From a Nordic point of view, the read is on who passes the price through. For a grid-electrical name copper is an input cost — grid-build volume is the tailwind, higher metal prices the margin headwind. For a Nordic miner the pass-through runs the other way and the Q3 release is where any realised-price benefit shows up. The level that would change the read on US tech concentration is an OpenAI or Anthropic disclosure — as they come toward market — showing cash revenue well below current implied pricing, combined with hyperscaler capex guidance cut by more than ten per cent year-over-year at the next reporting cycle.

Week ahead for a Nordic portfolio

The policy calendar is quiet. The next decision that matters is the European Central Bank on 29 October, with the deposit facility rate at 2.5% after the September hike. Norges Bank follows on 5 November at 4.5%, and the Riksbank meets on 12 November at 1.75%. None of these fall inside the next five business days.

The euro-area HICP reacceleration story carries forward. The September print came in at 3.8% year-on-year, up six-tenths from August at 3.2%. Italy printed 4.1%, Spain 5.0%, Germany 3.3%, France 3.4%, and the Netherlands 3.0%. That is the data flow that frames the 29 October ECB meeting.

Q3 reporting for Equinor, Maersk, Novo Nordisk and the big Nordic banks runs over the next two to four weeks. The two readings worth watching: whether a Nordic bank discloses US-dollar wholesale funding costs more than fifty basis points wider year-over-year, and whether Maersk's Asia-Europe contract rates have been re-priced into 2027 at a level that justifies today's share price. The read on Nordic consumer names exposed to China still waits for a volume-led guidance raise rather than a price-led one — nothing in this week's data suggests that is coming.

Nord Pool prices came off this week — Finland averaged 30 EUR/MWh, down from 49 the week before. The physical flows flipped, with Denmark's western zone net-importing from Germany on the week — a reversal from an export position the week prior. The consequence for a Finnish industrial consumer is a lower cost base entering Q4; the trigger worth watching is whether the eastern Nordic zones hold under 100 EUR/MWh across consecutive weeks, because that is what shows up in Nordic chemicals and paper margins before the Q4 release.

One structural item: the European Commission disbursed 1.24 billion euros to Ukraine this week specifically for drones and missiles. That continues EU-budget channels being used for lethal procurement, and it adds marginal demand to European defence and dual-use supply chains into 2027.

The direction is unchanged — defensive sector tilt into year-end, Nordic banks tested on Q3 funding and credit cost, grid-electrical names tested on margin pass-through, and the US index watched for whether the AI concentration can carry the weight.

This is a discussion, not investment advice.

This episode discusses investing in general. It is not investment advice, it does not describe anyone's positions, and nothing in it is a recommendation to buy or sell anything.

Sources