I think the author is speaking authoritatively about things they may be less familiar with, or where they really want to push a particular doomsday / degrowth agenda (the only prescription at the end the article is that we need to stop technological progress). This paragraph in particular caught my eye:
> Bah! Who needs copper anyway, when we have so much aluminum?!
> Have you thought about how aluminum is made? Well, by driving immense electric currents through carbon anodes made from petroleum coke (or coal-tar pitch) to turn molten alumina into pure metal via electrolysis. Two things to notice here. First, the necessary electricity (and the anodes) are usually made with fossil fuels, as “renewables” cannot provide the stable current and carbon atoms needed to make the process possible. Second, all that electricity, even if you generate it with nuclear reactors, have to be delivered via copper wires.
This seems to be trying to say that we can't make aluminum without copper, but that seems nonsensical. First, power can be delivered by wires made out of aluminum and indeed, it often is - I don't think that much of the transmission grid is copper. Second, the comparatively tiny amount of material needed for electrodes is a completely wacky argument. And renewables not being able to provide "the stable current" needed for smelting?
I'm not cherrypicking here, there's a lot of assertions of this type in the article. Essentially, everything is doomed and there's nothing we can do, because we're going to run out of copper. And fossil fuels. And there's absolutely nothing that can replace them, ever. And therefore, we shouldn't build AI datacenters? That's what it says...
> Second, all that electricity, even if you generate it with nuclear reactors, have to be delivered via copper wires.
This is indeed a massive red flag. You need conductors, but the material they are made of is pretty much irrelevant.
These days you'd have to search quite a bit to find not-ancient copper conductors in the larger electric grid. Aluminium might have a slightly higher resistance, but when you can just use a thicker wire it's almost always the more attractive choice.
If you don't even know that the grid mostly uses aluminium, you probably shouldn't be making big claims about what is and isn't possible with copper wiring.
Aluminum is actually a (far) superior conductor to copper per unit mass. It would be used on transmission lines even if it was the same price as copper, because the towers can be cheaper and farther apart. It's in increasing use in EVs due to the lower mass.
Copper is still used when the conductive density matters, like the windings of an electric motor. But if copper prices increase further, manufacturers will make sacrifices to efficiency and power density in order to save cost. And they'll figure out how to better balance the use of Al vs. Cu, perhaps using Cu only for the conductors closest to the core.
We also use copper for transformers, which are fairy "dumb" in their usual design. Solid-state transformers exist, which use much less copper, but are currently more expensive. They will no longer be more expensive if the price of copper goes up too much. And they'll probably get cheaper in the long run anyway, regardless of copper price, in the same way that switch mode power supplies have totally replaced linear supplies in the consumer space.
I've seen increasing use of copper in fairly mundane uses, like computer heat sinks, that used to be aluminum. The performance is a little better, but it won't be worthwhile if copper gets way more expensive. They'll just go back to aluminum, or use some other innovation (carbon heat spreaders, etc.) if price becomes an issue.
I'd suppose the fashion goes somewhat with the price of copper, though I haven't tracked it. The heatsinks themselves have gotten far larger as CPUs and GPUs have gotten more power hungry, not to mention RAM and SSDs. A material that's a good tradeoff at one scale isn't necessarily one at a different scale.
At any rate, one should expect many of these trades to go the way of Al if Cu gets more expensive (which it might not). Not all of them, but we'll probably see a bias towards physically larger systems in cases where space isn't at a premium. And also a bias towards active systems over passive, liquid cooling over air, and so on.
> First, no, power can be delivered by wires made out of aluminum and indeed, it often is, I don't think that much of the transmission grid is copper
Seconded, aluminum works just fine as a conductor. I’m pretty sure that all overhead utility distribution conductors are a steel core wrapped with aluminum conductors and air for insulation, and I’d bet that underground distribution conductors are also aluminum.
SER cable from the utility transformer secondary to your meter socket also uses aluminum conductors.
You usually need to go up a couple of sizes for aluminum vs copper (#1/0 Cu ~= #3/0 Al) but it depends on the specific ampacity.
> I’m pretty sure that all overhead utility distribution conductors are a steel core wrapped with aluminum conductors
That's the classical setup, but there's been some innovation. These days there's also stuff like aluminium alloys which don't need steel reinforcement, or aluminium reinforced with carbon fiber.
Thanks for the additional info, I’m interested in learning more technical details about the electrical grid, like conductor size/ampacity but it’s hard to dig up information about it. I work in commercial construction and the highest voltages I ever deal with are 4160V and sometimes 5kV and 15kV, so I know a bit about medium voltage equipment and conductors but I’d like to know more about the utility side.
Are you on a mobile device? Ampacity charts for aluminum conductor steel-reinforced cable are pretty easy to find but most of them appear to be PDFs, like this one:
I don't know about other countries but in Canada, I can think of a few aluminum smelting operations and they're all geolocated in close proximity to hydroelectric dams.
Other countries are very much the same. Almost always located near giant hydroelectric generation facilities. Brazil + Russia are two big ones that come to mind. Probably China too.
It's a double whammo, because aluminium smelting doesn't require stable current. Modern smelters can modulate their power significantly, with even multi-hour full shutdowns not being a huge problem.
Considering how energy-intensive it is, this means there is quite a big future for using aluminium smelting to soak up dirt-cheap excess renewable energy. Renewables not being stable has suddenly become a feature.
>This seems to be trying to say that we can't make aluminum without copper, but that seems nonsensical.
The far better argument is that, if it were simple to replace copper with aluminum, this would create a ceiling on the price of copper. However, this hasn't happened. Many applications of copper can theoretically be replaced by copper, but in practice the reactivity and thermal performance issues of aluminum can be challenging. Aluminum wiring in homes, for example, has a very bad reputation.
This isn't fatal, but it is a problem. And if society doesn't plan for it, it could become a more painful problem.
> Aluminum wiring in homes, for example, has a very bad reputation.
It has an undeserved bad reputation now, but it deserved the bad rep back in the 60s and 70s. The problem wasn’t solely the aluminum conductors themselves, it was also the terminals on wiring devices. The material the terminals and screws were made out of worked fine with copper, but the thermal expansion profile did not work well with aluminum conductors. That caused arcing and fires, so the wiring device manufacturers figured out a material that works well with both copper and aluminum for wiring device terminations. Wire manufacturers also made changes to ensure better terminations. If you look at the terminals of a light switch or receptacle, it will say Cu/Al on it, signifying it is suitable for use with either type of conductor.
This was solved 50 years ago, it’s similar to being scared of flying on a modern jetliner
because the De Havilland Comet ripped itself apart in the 1950s due to the engineers not understanding the stresses from repeat pressurization cycles.
For existing installations of pre-1972 wire, you can buy splicing devices (similar to a WAGO lever nut) that connect to the aluminum conductors inside the box and allow you to connect a copper pigtail to the wiring device, you also have to use an anti-oxidant grease to prevent oxidation.
That being said, I’d still wire a house with copper because you can use #14 Cu for a 15A circuit but you need #12 Al for the same circuit, the NEC does not allow use of #14 Al romex.
> In North American residential construction, aluminum wire was used for wiring entire houses for a short time from the 1960s to the mid-1970s during a period of high copper prices. Electrical devices (outlets, switches, lighting, fans, etc.) at the time were not designed with the particular properties of the aluminum wire being used in mind, and there were some issues related to the properties of the wire itself, making the installations with aluminum wire much more susceptible to problems. Revised manufacturing standards for both the wire and the devices were developed to reduce the problems. Existing homes with this older aluminum wiring used in branch circuits present a potential fire hazard.
I live in a home built after 2000 that had aluminum wires run to its heat pump. A few years back coolant leak from the heat pump lead to huge electric usage before the aluminum wiring lit on fire and shorted itself out. Since repaired, but was told at the time original installer didn't correctly do the aluminum grease on the exposed wire parts.
That said I think the wiring there is still thick aluminum.
Not sure I really have a point - all things equal I'd prefer copper, but it seems like aluminum can be fine when done right too - just riskier when done to the quick and dirty homebuilder standard.
> You do need constant, reliable power, as even a brief interruption makes a huge mess when the aluminum/slag freezes in the processor
On the other hand: it has a gigantic thermal mass. Combine this with the energy requires to melt it, and you end up with molten aluminium being trucked over our highways [0]. A brief interruption isn't a big deal when it takes ages to solidify.
Iceland is a tiny country with unusual amounts of energy. Not all renewable sources are the same -- hydropower is fairly reliable too, for example -- but Iceland is just not a useful example for the whole world. The largest geothermal plant in the world by far is in California, but it's a small portion of our total energy use so no one cares. https://en.wikipedia.org/wiki/The_Geysers
You can locate an aluminum plant pretty much anywhere you want, as the energy required to make aluminum is large compared to the cost of mining/shipping bauxite. This solves the main problem with geothermal, which is that it's in random locations around the world that don't necessarily have many people living there.
Any place with significant volcanic activity (e.g. Hawaii) could probably do geothermal power if they wanted to.
> Even though the industry would be willing to pay top dollar for each pound of metal delivered, there is simply not much more to be found. Copper bearing formations are not popping up at random, and there is no point in drilling various spots on Earth prospecting for deposits, either. The major formations have already been discovered, and thus the ever increasing investment spent on locating more copper simply does not produce a return.
How do we "know" there isn't any major formations we haven't found yet? I find it hard to believe we've prospected every possible area.. or are deposits more predictable than it seems?
The mining majors, BHP, Rio Tinto, et al have petabytes of surface geochemistry, samples, near surface magnetic maps that penetrate into the crust, 3D seismic maps, drill cores, technical reports on every mine ever, surrounding geology, and good models on where economic feasible (at particular price points) amounts of desirable metals can be found.
For example, there are only so many places significant masses of porphyry copper deposit will be found (although these aren't the only types of copper deposit).
For people interested in subscribing, there are databases such as the S&P portal that scratch some of that industry knowledge.
although they seem to have backed off from a public page about the GIS portal to the mining databases they purchased.
So; pretty much most areas have been scratched - Antartica is still open, the Artic has possibilities .. but should we.
There are known untapped masses of copper, eg: in the US there's a mass that will take 64 years to mine .. that's on Apache land so, you know, it'll be US history all over again poking that one.
For contrast (and with no AI involved) the dot points from a Sept 2024 BHP copper report are roughly:
* Total global copper demand has grown at a 3.1% compound annual growth rate (CAGR) over the last 75 years – but this growth rate has been slowing. It was only 1.9% over the 15 years to 2021. Looking to 2035, however, we expect this growth rate to jump back to 2.6% annually.
Considering global demand expectations, electrification trends, etc. this moves to:
* Putting all these levers together, we project global copper demand to grow by around 70% to over 50 Mt per annum by 2050 – an average growth rate of 2% per year.
On the supply side, considering recycling, the aluminium usage transition point, etc:
* We estimate that the world will need about 10 Mtpa new mined copper supply in the next 10 years.
To get this ...
* Copper reserves and production are concentrated in Latin America, Australia and Africa.
* We expect supply growth over the next 10 years to be dominated by the same regions – Latin America Africa and Asia Pacific – with Africa having the highest growth rate (albeit off a much lower base than Latin America), and Latin America continuing to make the most significant contribution in absolute terms.
Existing mines are falling off (they always are!):
* Currently operating copper mines are expected to provide more than half of the copper required to meet future global demand over the next decade. Even so, we estimate existing mines to be producing around 15% less copper in 2035 than they do today.
* We expect between one-third and one-half of global copper supply to face grade decline and ageing challenges over the next decade, which will drive increased unit costs and the requirement for capital reinvestment. While an incredible orebody can make a big difference, many older operations move up the cost curve as they progress through their life cycle. Given the strong demand signals, however, we expect the industry to vigorously pursue options to extend the life of these copper mines.
Brownfield developments will expand old mines; the practice of processing satellite deposits ignored early on.
Truly new mines:
* Greenfield projects continue to attract significant excitement and interest from developers and investors. They can avoid the challenges of aging facilities and grade decline and can unlock large and higher-grade copper deposits, develop new frontiers, and allow for the application of technology advances without the challenge of retrofitting.
* But they also have potentially even greater challenges to brownfield developments, such as long lead times with environmental and social concerns needing to be navigated for the first time, and uncertainties associated with new jurisdictions or regions. And not all problems can be solved with money. For some projects, it is not a question of investability, but of executability.
* The current pipeline of ‘all possible’ greenfield deposits are generally at the higher-difficulty end of the spectrum – and many are experiencing delays. When we investigated a selection of today’s 30 largest (by expected production volume) undeveloped greenfield projects, we found that analysts (ourselves included) had continually moved the forecast supply stack out in time. We expect these projects to contribute around 5 Mtpa of copper by 2035, or 14% of total possible supply.
There are other similar reports from the IEA and other majors.
This is the literal ever ongoing dance of extraction, exploration, and development.
Every cent that goes to exploration is whinged over and penny pinched - it's all buckets of money out the door with no return other than bluesky hope.
As prospects appear and are firmed out, speculative money starts to flow, once Technical Feasibility Reports are in the works a frenzy begins, the bets may or may not pan out when such reports land in a Prospectus seeking forward capital investment.
Yet another author who does not understand the meaning of reserves vs resources. Reserves are are deposits that can be recovered at cost at current market price. Increase the price and miraculously, reserves increase. Like with most commodities, price spikes are followed by price drops as supply catches up. Remember when we were running out of lithium? Prices are down 90% from that peak.
> Bah! Who needs copper anyway, when we have so much aluminum?! > Have you thought about how aluminum is made? Well, by driving immense electric currents through carbon anodes made from petroleum coke (or coal-tar pitch) to turn molten alumina into pure metal via electrolysis. Two things to notice here. First, the necessary electricity (and the anodes) are usually made with fossil fuels, as “renewables” cannot provide the stable current and carbon atoms needed to make the process possible. Second, all that electricity, even if you generate it with nuclear reactors, have to be delivered via copper wires.
This seems to be trying to say that we can't make aluminum without copper, but that seems nonsensical. First, power can be delivered by wires made out of aluminum and indeed, it often is - I don't think that much of the transmission grid is copper. Second, the comparatively tiny amount of material needed for electrodes is a completely wacky argument. And renewables not being able to provide "the stable current" needed for smelting?
I'm not cherrypicking here, there's a lot of assertions of this type in the article. Essentially, everything is doomed and there's nothing we can do, because we're going to run out of copper. And fossil fuels. And there's absolutely nothing that can replace them, ever. And therefore, we shouldn't build AI datacenters? That's what it says...
reply