800,000 tons of mud probably just made electronics a little more expensive In recent years technology buyers have endured hardware price rises due to a pandemic and its impact on supply chains, the global wave of inflation that followed, tariffs, and surging demand for AI technologies that allowed vendors to charge higher prices. Now, 800,000 tons of mud has pushed copper prices higher. The mud slid …
There's not THAT much copper in electronics. A 19 cent chip has a small fraction of a cent in copper leadframe. A $200 CPU may have a whole penny. Yes PCBs have multiple layers of copper but engineered to a minimum.
But think of people wiring or re-wiring houses! My state has a housing deficit and wiring was a big chunk of my last project. But 3%??? Over a few years the price of wire doubled, another 3 cents is just annoying not prohibitive.
Copper re-cycling is omnipresent. I remember when the US was throwing brass cartridges in the swamp of Vietnam driving copper prices up. Since then copper recycles itself: copper ground wires on power poles or copper pipe in vacant houses quickly gets stripped and sold. I upped my ground-rods because I can't trust taking ground from the power company when every addict has wire cutters.
There will be businesses that specialize in cannibalizing electrical wiring and everything else of value from half built or older obsolete datacenters once the AI bubble pops. When the fiber bubble popped they left it all in the ground because there was no recycling value and whoever bought it for pennies on the dollar was counting on it being useful 10 or 20 years later. That won't be the case with datacenters, they're gonna gut those.
The owners are shell companies, so OpenAI, Meta, and so forth will likely just abandon it once they strip it bare of anything worth taking, and leave the concrete edifice as a problem for the city/county government to deal with. Some will try and sue, but Meta has lawyers that individually make 10x more than the whole county's legal staff so it won't be a fair fight.
The electrical grid wiring is mostly steel, but once you reach the datacenter site all the wiring is copper. And as the power density of racks grows so grows the amount/thickness of wiring. Nvidia projected that a single AI rack may require a megawatt by the end of the decade.
That is not going to be easy to deliver - normally a datacenter has three phase 208v/240v to the racks, at least that's what all the ones I ever worked with had. If you have 50 amps to each phase that's 36kw. Not too long ago that 36kw was for an entire row, so you had one wire split off in junction boxes under the floor with a couple pigtails going to each rack's PDUs (I know some use overhead copper busbars but I never saw a datacenter using that personally) But it grew and grew until you needed that much power for a single rack - which means a length of wire from the distribution panel running all the way to each rack (or the busbars I guess) WAY more copper! Now there are racks using 100kw or so, but I never worked with those so I'm not exactly sure how they deliver that power. But however they do it they're gonna need more copper!
So how in the world will they deliver 1000kw? I'm gonna guess they jump to 7200v, because there is a lot of electrical gear made for that voltage because it is what is traveling on overhead or underground lines in many neighborhoods. That voltage allows them to stick with 50A, and avoid using too much copper for thicker hard to work with wiring/busbars. I'm not even going to think about the plumbing required to move a megawatt hour of heat every hour out of each and every rack, but there could be some copper involved here and there in that mess.
7200v is incidentally gonna make the electrical work a lot more expensive, because typical facility voltages only go up to 480v, beyond that electricians need additional licensing and safety procedures due to arcing etc. and that means $$$! So if you are or know an electrician, getting trained up on 7200v could pay off quite well if a datacenter gets built in your area.
The electrical grid wiring is mostly steel,
The overhead transmission grid wiring is mostly steel by weight. By volume, it's majority Aluminium.
Underground cables don't need the steel core, so they have a solid aluminium core.
The residential grid may be copper in your area, but new lines are steel/aluminium like the transmission grid.
That only works if the ring isn't loaded everywhere so the total load is only 30A. That concept works well in a domestic environment but not in a data centre. The other thing ring mains have to have is a fuse in every plug. You can't rely on the big breaker in the consumer unit.
>> a fuse in every plug --- >You say that like it's a bad thing?
Who replaces the fuse? The same consumer who just blew it. When they get half a clue, they upsize the fuse (the whole plug if that's what it takes).
FWIW, the standard US/Can lamp-cord is intended to reliably blow a 20A breaker without setting fire to the carpet. The older (pre-1960?) lamp-cord did not and that problem was addressed when standard circuits shifted from 15A to 20A.
Ring-main circuits are vastly better for specific load distributions. In real world loading the advantage is slim. I wish it were otherwise: wiring a whole house octopus style is a LOT of heavy copper to drag upstairs or thru walls. I even looked at it for my garage loft where ring-main was a good physical fit (no internal walls, few large openings). But it was hardly-any savings, and I could easily overload it. I think it was a post-WWII stopgap for a time when copper was very dear and very few large appliances were on the market.
It really isn't. I get what he's saying about changing fuses etc...but here in Blighty it's pretty easy to find someone in your immediate area that can change a fuse in a plug. Just shout around the room.
The only person I wouldn't trust to change a fuse is an old guy I used to help out occasionally...he was a radar engineer in WW2...whenever he blew a fuse, he "replaced" it by shortening the wire to plug, and using some of the "spare" wire to short the fuse compartment by soldering it in...well his form of soldering...which was heating up a metal stick with a blowtorch until it was glowing then dribbling solder in to the cavity off it. Dude didn't fuck around...he used to call me around because his wife was sick of him burning holes in the carpet with his iron. He died at the age of 93, dude was an absolute badass...and I'm amazed he lived that long, or maybe he didn't and lived a series of short lives shocking himself to death then back to life again...he was the kind of engineer that licked his fingers and touched the wires to see if they were live. Truly the Johnny Knoxville of electricity. Working on his "projects" that he'd been maintaining for nearly 2/3rds of a century hardened me up a bit I think...you couldn't touch half the stuff he built without getting a bang off it at some point...man the shit he did, I have stories for days...boiling his tea with a bare immersion heater, heating soup in the tin with a Bunsen burner...man was the 18+ Wes Craven version of Doc Brown...he was a professor for decades and I once asked him "based on your safety record, how many of your former students have died" to which he replied "oh none for a while"
That concept works well in a domestic environment but not in a data centre
True, but the OP did say "think of people wiring or re-wiring houses". It's an ideal setup when you want the flexibility of many relatively high-power outlets but few in use at any one time and a low overall load, which is typical in a domestic environment.
The other thing ring mains have to have is a fuse in every plug. You can't rely on the big breaker in the consumer unit.
They are there to protect different things. The fuse in the plug protects the low-current appliance cable, the breaker on the consumer unit protects the ring.
Apropos to copper recycling:
Bangor man stealing copper pipes caused house explosion, court records say
https://www.bangordailynews.com/2025/09/26/bangor/bangor-police-courts/bangor-fire-copper-theft/
because this paper's CDN sometimes refuses out-of-town surfers I'll skim:
.....man is accused of causing a fire in March while trying to steal copper pipes from a vacant house. An explosion and subsequent fire erupted in a vacant home around 5:45 a.m..... ....investigators found people were removing copper pipes from the basement,... There was a small pile of cut copper pipes that were roughly a foot long..... One cut pipe was going to the gas furnace, .....That line likely explained the gases that caused the fire, the affidavit said..... was using the light from the drill to see and a spark from that may have caused the explosion, she said. ........in bed with a blanket covering his face, the affidavit said. Once ....lowered the blanket, police saw he had serious burns to his face, the affidavit said. A burned sweatshirt, a blue lighter torch and pieces of tin ceiling matching the Fourth Street house were found ....
How much of that is now buried under the mudpile along with the poor sods known or presumed to have been killed... I know they got a brief mention towards the top of the article, but I do feel a touch uneasy seeing just how focussed the article and subsequent comments have been on the material/financial impacts of this incident.
Demand is going have the biggest impact on copper prices.
Trumps idiotic tiny-willy waving tariffs are probably next (yay Republicans, fscking it up again)
Residual impacts of the COVID pandemic are probably third most significant - but, by now, probably not all that great.
The mud slide, in comparison with the top two points, is a rounding error. Significant for the owner of the mine but, globally, not so much.
However, I should make it clear, that this opinion is unevidenced and based on gut feeling. But hey! This is the 21st century. We don’t need experts where we’re going!
A 3% price increase due to a 3% reduction in supply is much less than I would have expected. That indicates there isn't a lot of pressure on copper prices at the moment.
One of the big drivers for copper consumption was China's building boom, but they are still "carefully deflating" that bubble so the demand for copper in construction in China is far less than it was a decade ago when prices really spiked.
"Significant for the owner of the mine but, globally, not so much."
On the other hand, it amply demonstrates the broken stock market. The short term investors have caused a "panic", undervaluing the "victim" and overvaluing the rest of the market because a very short term blip in production, Investments in things like mining raw materials are long term investments and a blip like this is just par for the course. Odds are that when the mine goes back into production, possibly with the addition of new/replacement equipment, in 6-12 months they will be more productive and will then be overvalued by the short term investors because the lost production/profits will have been forgotten and only the huge increase in production from near zero back to "normal" will be seen as a giant leap in profits.
