Global Foundries was originally the manufacturing arm of AMD, and was then spun off as a separate company, acquired the semiconductor operations of AMD, and was bought by the UAE, not necessarily in that order.
The company is planning to invest $12 billion over the next decade, mostly on specialised production like gallium nitride and expanding bulk processes on 12nm.
Is this money well spent? Probably not; if it were, Global Foundries wouldn't need the money. Is it money better spent than almost anything the US federal government does? Maybe.
@Marakus@Aether it wasn't *that* long ago that local manufacturing and sourcing was the norm virtually everywhere for everything. sure, people were importing niche goods they didn't have locally but they weren't buying industrial feedstocks from the other side of the planet
anyone, anywhere should be able to make silicon crystals of sufficient purity with the right expertise and enough energy to pump into the process. you can find relatively clean deposits of silicates literally anywhere on earth
@deprecated_ii When it comes to strategic resources like Chips, pretty much every country needs a localised manufacturer. Trouble is that Globo Homo is against any such thought
@Aether@deprecated_ii Chips manufacturing require long supply chain. Even finding non-smocking personal might be too much for some countries, never mind reliable supply of pure silicone wafers.
@Aether every country should be subsidizing strategically important industries on their own soil
but they should also be *extremely* careful who they let run such things, and nail said people to the wall (literally) when they start fucking around, skimming money, shirking their duty, etc
of course we're not doing that second part. this will just be another giant grift for insiders to get rich off of
@Marakus@sapphire@Aether You think it's harder to get high purity silicon from beach sand than it is to extract trace metals from hard rock deposits a couple thousand feet underground? Because we do the latter every day, all over the world.
It may be slightly more expensive to use non-ideal sources, but it's rank bullshit to claim it can't be done pretty easily. The companies who claim they need some super special feedstock are just snake oil salesmen.
@pyrate@Aether@Marakus@sapphire I've never dug into it but I wouldn't be surprised if the sand they want to use just naturally has some of the right impurities so it's a little bit cheaper and easier to dope.
At global economies of scale, the difference might mean the CEO gets to add another deck to his yacht, which is after all incredibly important. Critical, even.
In fairness. I would imagine regular beach sand to contain a massive assortment of impurities that would make it too difficult to process. Alas they have to use mines.
Or perhaps a more intelligent strategy would be greater precision recovery processes from existing e-waste.
@deprecated_ii@Marakus >it's harder to get high purity silicon from beach sand As far as I can tell, semiconductor grade silicon is almost never made from beach sand due to the extra expenses in having to purify such a low quality feedstock, but use cases that tolerate 96% purity silicon work fine with beach sand.
From what I can tell, typically quartzite is carbothermically reacted with highly pure coke to leave 99% pure silicon and then there's multi-step purification processes that include conversion to silicon tetrachloride or trichlorosilane and purification to like 99.9999999% purity.
Even with 7 9's of purity, one still needs to grow defect-free silicon crystals, cut them into wafers and then dope those wafers so the silicon becomes a semiconductor rather than remaining a insulator as pure silicon tends to be.
@pyrate@Aether@Marakus@deprecated_ii@sapphire Some teenager is probably going to figure out how to refine beach and into silicon in his garage using a burr grinder and a hot plate. TBH this sounds be a good thing to try to push Nile Red to do.
@pyrate@Aether@Marakus@sapphire I would expect the right beach or alluvial sands to be quite pure actually, from a mineralogical point of view. Not to get into a geology lecture but it has to do with what minerals persist well in surface conditions and which tend to be weathered into other things, and quartz -- SiO2 -- is extremely persistent.
Now if they're mining hydrothermal quartz veins or something maybe that's even better. But that's also fucking crazy unless it's waste rock from mining other things, which it could be, because hydrothermal quartz is associated with precious metal bearing sulphides.
@pyrate@Aether@Marakus@deprecated_ii@sapphire I had no idea that was a thing but based if so. I wonder if street sweepers companies have ever looked into vacuuming stuff for minerals.
It wouldn't be the first time I've seen aspiring and poor White science types doing it. Many collect dust off road ways during the wee hours to filter out and refine platinum.
@pyrate The thing is, e-waste can't really be recycled.
e-waste "Recycling" typically consists of one or a few of the following; - Desoldering off old components and electrolyte caps etc to sell as new - typically only a few components per board can be recovered. - Sanding off coatings etc and dumping boards into vats of mercury to extract the gold. - Grinding down whole computers and extracting the metals as scrap.
Computers don't really contain much silicon, as typically the die etc in the CPU is very small, while the rest of the computer is mostly metal, with PCB's usually made out of sheets of copper and fibreglass or paper impregnated with a phenol formaldehyde resin .
Even if you have a bunch of CPUs to grind down, such silicon is of poor purity as it's doped and has a bunch of gates etched into it.
Once "recycling" is done, anything that remains is either dumped wherever or into landfill.
Companies really love shilling the idea of "recycle your old computers and buy our new ones that are even more proprietary, to save energy", but really when you run the numbers as to how much energy it takes to manufacture a computer, it's more energy efficient to run a old computer for 30 years than to buy a new one.
@deprecated_ii You do indeed need 7-10 nines purity to make a CPU at modern tiny process sizes.
At current process sizes, any impurities at all kills yields.
Older process sizes that used gates hundreds of micrometers in diameter work just fine out of silicon at lower purities, which is how the industry was bootstrapped.
Non-microprocessor applications tend to be more tolerant of impurities and crystal dislocations - solar panels work fine when made out of polycrystalline silicon for example.
@Marakus@sapphire@Aether There's no way you actually need super dooper 10 nines purity to make *a computer chip*. Maybe for cutting edge stuff that's 10000x more powerful than most applications require.
We never would've bootstrapped the industry in the first place otherwise.
@Marakus@Aether@deprecated_ii yes. There are established electrochemical processes that take sand to 98.5% silicon for industrial use and an established process to purify and crystalize* that silicon into boules for production.
Do you seriously think the Chinese are mining semiconductor grade silicon from the fucking ground lmao
@sapphire@Aether@deprecated_ii 99% is about million times more impure, than what is required for chip production. Getting fucking sand is not a problem, purifying it to 99.999999% is.
@deprecated_ii@sapphire@Aether No, its difficult because you can only sell it to chip manufacturers, and you cant produce it on small scale. These companies make hundreds of billions, its all or nothing. Its not fucking sand, its pure 99.99999% silicone. With that logic "but its just sand, its cheap" you should be able to breathe water, since its mostly oxygen.
Quite possibly. Last neighbor I lived next to had a sweeper truck he would go do parking lots and little private neighborhoods at night. Never spoke with him, but now I really wonder.