I'm in charge of sourcing components for multiple companies right now. I believe they could double (100%) prices and companies would pay it. There's chips that I'm sourcing from obscure places in China just to get the part and paying whatever price they ask.
Designers are getting overwhelmed because they're having to redesign boards for components that are available, only to realize by the time they finish the changes, the part they redesigned for is also out of stock.
There's been some talk of a Chip-induced cold war, and I would say that's not crazy given the market conditions. I also love the guy who posts here all the time saying he works for the company that makes the machines that makes the chips and they can't make more machines because they don't have the chips that go in them. XD
Concur. We are currently paying a chip broker $6.30 for an Infineon chip we typically buy by the 100,000s for $0.47. Cuz without them, our customers cannot sell their $10,000,000 product.
24 volt digital output drivers.
We're also getting crushed on Xilinx FPGAs, certain Cypress CPUs, AD MUXes, Relay switches... It's something new every week.
Venor Vinge's The Children of the Sky is hard sci-fi touches on this issue of how to bootstrap a civilization to the point where the technology necessary to restore a highly advanced spaceship becomes available. Humanity will need this skill if we are ever able to venture out of our solar system.
I've played enough survival games to know that all you need to do is fashion a pickaxe out of stone and wood, use it to mine some metal, build a forge, and everything just snowballs from there
For instance, knowing that Corn doesn't supply absorbable niacin without nixtamalizaion, and diets dependent on corn tend to have large bouts of pellagra, is something that took hundreds of years to work out.
Walking into a corn fed post-apocalyptic future knowing that soaking your corn in ash water and hulling it prevents pellagra and improves the viability of your group is some handy knowledge.
This is also why rebuilding a web page after deleting it is so much faster the second time around. You've solved problems on the first go round that weren't apparent when you started.
> This is also why rebuilding a web page after deleting it is so much faster the second time around.
Counter-intuitively though, while a lot of people think rebuilding software from scratch will be a lot faster, in practice it's not because people tend to have to re-invent and re-discover a lot of things, and bolt a lot of other stuff on top as well. Plus over-engineering, plus analysis paralysis, etc.
So making a carbon copy of e.g. a website: sure, but for a lot of software it doesn't work like that.
That said, I do believe rebuilding software will work out better in the long run because learning from the mistakes of the past. It'll be interesting to see what e.g. Google's Fuchsia will turn out to be, the first big / sponsored OS made from scratch in a while.
And Firefox / Rust is another interesting one; it was a massive investment and risk, but I think it's paying off.
I have the same experience in general programming. You start doing something but then you run into a dead end. Then you look for other routes to your destination and find them. If you do it the second time you can avoid those dead-ends, which you didn't even know about when you started.
I think this means we should keep log-books like ship-captains do so we can avoid treacherous waters on subsequent trips.
It's a nightmare. A broker in HK recently quoted me $71/ea for an industrial temp rated 10/100 Ethernet PHY (KSZ8081MNXIA) ... That is a roughly 50x markup. We're making a design change instead. I am now buying qty 500-1000 of any new part that is getting designed in, before the design is even started.
Our Analog Devices sales rep told me to expect another 12-18 months of everything being completely fucked. That was based on a conversation he recently had with TSMC. It's going to be super fun.
Oh yeah, I've told my manager it's only a matter of time until we get burned. So far we have had decent luck with Heisener and Win-Source but prices keep climbing.
Well.. we were just looking at Heisener for a Cypress chip, but can't decide whether to pull the trigger. Have you placed an order and received the product from them?
Yes, we successfully purchased a small quantity of Kinetis microcontrollers which were supposedly tested prior to shipping, those are going into assemblies right now so we'll see if they actually work.
Heisener is just a middleman so every transaction is different unfortunately.
> by the time they finish the changes, the part they redesigned for is also out of stock
One of the things my last employer did to get around availability problems was to buy a vendor's entire supply of a chip we could use and then redesigning the product to use it. IIRC, the supply was projected to be enough to last a year. Hopefully, a year from now it will have proven to be a good decision :-)
Similarly, my company just redesigned a product after running out of an IC that's been on order for many months, still without an estimate on delivery by the distributor.
We redesigned it to use another chip we had in-house, intended for another product, but that product sells for much less and is in less demand, so...
We initially thought we could move pins around in firmware, and the parts have the same footprint, so it seemed possible, but they were different enough that we had to rev the PCB. Now we're waiting for PCBs and crossing our fingers that they will test out.
Um, I just did the same thing the GP did (buy up all the stock) for a product in development but one of our older products is getting hit by the same issue described by the parent!
Bosch recently opened their own chip fab in Dresden, primary focus for now is on automotive chips and those required for their in-house tools and appliances:
This makes a lot of sense as Bosch make so many automotive parts. In Europe, it doesn't matter whether you buy a BMW, a Peugeot or a Volvo, under the hood they are all using Bosch electronics to control the engine and many other parts of the car.
"service economy" was always a meme being sold to the gullible masses, manufacturing capacity is the only thing that matters for an economy. Manufacturing has always printed the money
It amazes me how the US allowed itself to be nearly deindustrialized after manufacturing effectively won them WWII and global dominance. Now the US can't even make ventilators or face masks
US Manufacturing is still going. In fact, we manufacture more than ever. Folks conflate manufacturing with manufacturing _jobs_. Our manufacturing is more efficient then ever which is why you constantly hear this false refrain that US manufacturing is down or gone.
It seems to me that if you can make a product of a given quality very cheaply, that probably indicates that you're very good at making it.
The fact that I can't get competitive production quotes in the USA for most products, like electric motors, injection molded plastic parts, printed circuit boards, stamped metal parts, castings, etc, suggests that the US is also not particularly efficient at making those parts. There was a time when the cheapest place to go for such things in production volumes was actually the US.
The price quotes you are getting also include the opportunity costs of saying no to other people wanting to pay for that manufacturing capacity.
The people I know that are running manufacturing aren’t exactly hurting for business. Now is just not the time to lock in high quantities of low-cost stuff.
This article is about TSMC raising prices for what little capacity they have unallocated. We can all agree that they are probably the best in the world at what they do, yet there is no way you will get “competitive production quotes” from them for commodity chips. Jim Bob’s Plastic and Stuff out in Fort Wayne is no different. Now, Jim Bob is probably going to cut you a deal way sooner than TSMC is, but I know of a certain mega-big car company that shut down an engine plant for a week because Jim Bob said last years prices aren’t this years prices - and you probably aren’t as big as mega-big car company.
people always point to this dumb metric, "we're producing more than ever" compared to ourselves, while ignoring that the number in comparison to the rest of the world has dropped massively. Our share of global manufacturing is tiny compared to what it used to be
I've worked for manufacturing companies my entire 30+ year career. It's not a dumb metric. It's usually pulled out in response to the silly complaint that "the US doesn't manufacture anything anymore."
Sure, our share of global manufacturing is smaller, but that's hardly surprising when many more countries up their manufacturing game. A more interesting metric would be US percentage of total worldwide manufacturing as a function of sophistication: I'd rather be building complex medical instruments than toasters.
Most toasters suck. I have to buy a new one every other year or so, for some stupid electrical or mechanical reason. Right now, for instance, the latch on our current toaster -- the one that latches on to the handle you press down to lower the bread and close the circuit to start the heating elements -- isn't latching any more. Fucking low-grade crap! I mean, I'm not asking for a self-lowering magically-correct-degree-of-toasting toaster like in the old TC video -- I just want a stupid mechanical latch in a product I bought a couple of years ago to keep stupidly mechanically latching for six or ten more years. That is not too much to ask.
So, stick to your complex medical instruments. Or, if you switch to toasters, please don't build the same shitty crap as everyone else.
>I'd rather be building complex medical instruments than toasters.
Why? In what way is it better to have a handful of elite workers creating a handful of units sold to a handful of consumers instead of having many ordinary workers producing many units for many consumers?
What happens? Well I suppose the economy collapses and the proletariat rise up and usurp power, or something fantastical.
What’ll more likely happen is we’ll further automate repetitive low skill tasks, and / or start making more durable goods, or otherwise ingenuity our way forward.
Post-industrialized countries make it impossible to for those countries arriving late to the party to advance any further, by way of global environmental agreements and glass bead "carbon credit" payments, thereby ensuring a permanent pool of slave labor.
Eventually it becomes more profitable to automate. See: Sony's automated PS4/5 factory. Formerly these would've been assembled in some Foxconn plant by tens of thousands of low wage workers.
I have a sneaking suspicion that's fundamentally impossible, as some of the prerequisites for developed economies (in the modern, not-industrialization sense) seem to require predatory exploitation of developing economies. And the global inequality of capital ensures there's a power disparity sufficient to enforce it for the former, to the detriment of the latter.
But I guess it's really a question about post-scarcity economics, and whether they're possible?
Because there won't be "many ordinary workers" producing those toasters. Low margin products like that will have their manufacture automated to the n'th degree and they'll be built by "a handful of elite workers" supervising the automation.
Or, you know, they'll send it to the lowest-cost labor supplier they can find to build by hand and that won't be in the US!
That's because the rest of the world is no longer a bombed-out postwar wasteland or its near-enslaved colonies as it was in 01945. "Producing more than ever" is what you needed last year when you couldn't get N95 masks or ventilators, not "producing more than the rest of the world". But something went wrong and the masks and ventilators were not forthcoming, something that looks an awful lot like an absolute decline in manufacturing abilities.
Other people said this above, but I will add my own anecdote I heard from a factory owner, that he had the ability to retool immediately back then to make medical masks, but between the uncertainty around whether people would keep buying them and at what cost, he was unsure it was worth the risk of switching away from a profitable business to something that would take less ability to make and less certain profits. I don’t know what he ended up choosing to do. But sounds like ‘not forthcoming’ is also possibly an indicator that skills, and the demand for them, is high—and not in decline.
> Our share of global manufacturing is tiny compared to what it used to be
You mean compared to a time when most developed countries' production capacity was significantly limited because they had been bombed out in WW2 and developing countries were a long way away from being competitive?
tl;dr if you focus on Intel's monopoly money-printer years, multiply it by a hedonic adjustment due to Moore's law, and mix it in with the rest of manufacturing it's enough to hide de-industrialization in your statistics.
> But it turns out that Trump’s story of US manufacturing decline was much closer to being right than the story of technological progress being spun in Washington, New York, and Cambridge.
Claiming Trump was really right, is usually sign of a writer trying to get people to read a bunch of opinions that don't match reality presented as "bold independent" stance
> Worse than the Great Depression: America’s manufacturing jobs implosion
So they're straight up going to max bullshit mode?
And they say if we ignore one of the most productive manufacturing sectors (computing) the data looks worse, Is that at all surprising?
US is smart enough to retain the cream of the manufacturers. I think some smaller western countries are really in the danger of a complete de-industrialization.
It's not for lack of trying. It's hard to compete on the global stage and it takes more than wishful thinking to maintain a competitive edge in a world full of people trying to top you.
Not really, it's usually the least profitable part of the supply chain.
That American businesses would gladly hand over their profits to poorer countries is irrational at face value.
TSMC is more than a manufacturer, they are a provider of very sophisticated in-demand tech for which the manufacturing process itself - is high tech, high stakes, high barriers to entry.
They are a little bit like Boeing in that the 'making' part is really hard. Less so like Flextronics.
Depending on how countries react to this, we could very well be in the opposite situation in 3 years, where we have a vast glut of chips and parts of every kind.
TSMC is not going to charge their customers 'prevailing market value' i.e. they could spot price at 5x their current prices, because it'd be gauging and the supply change would react long term to that 'bad behaviour' by TSMC. Big supply chain relationships are longer term so those things matter.
Because of the huge players involved with enough cash do influence (i.e. Auto, Apple, Samgsung) maybe we're in for a strategic alignment beyond just a fab shop in Texas.
One big caveat to all of this is that though manufacturing itself is low profit, a lot of things around that can become profitable. Making moulds, adjusting designs, managing supply chains, optimizing designs etc. - all of this becomes 'know how' that moves over to the manufacturing side and if they are smart about it, they'll leverage it.
Also, manufacturers can possibly develop some market power which they can try to flex as well.
Hopefully we see a shift to more high quality intelligent manufacturing in the Western world by leveraging tech, robots etc. and convincing consumers to pay higher prices for some things.
Service economy doesn't imply deindustrialization. The provision of a service can be highly capital-intensive.
>>Now the US can't even make ventilators or face masks
The US led the world in the speed of its vaccination rollout. As for manufacturing, its world-leading biopharmaceutical infrastructure allowed it to produce the most shots of the most effective covid19 vaccines, of any country in the world.
"Manufacturing economy" was a meme sold to gullible nostalgic masses
Work is work, money is money. There's nothing fundamentally more important about manufacturing then other types of jobs. The US still has tons of manufacturing but unsurprisingly its on the backs of much fewer jobs, many of which require more education and specialization then manufacturing economy of old used to supply
> There's nothing fundamentally more important about manufacturing then other types of jobs.
the idea that manufacturing is the only work that matters is especially bizarre on a social media website for tech workers.
microsoft or oracle or any number of other software products absolutely deliver value to customers just as much as a machine stamping out parts, those companies couldn't work without that software just like they can't work without an assembly line or a delivery truck or whatever.
Luckily you can buy computers from other countries
Same with food.
It's not anything special.
A way to earn a living/have an economy as a country is as good as any other.
Moldex masks are made in the US. Their unique construction also protects them somewhat from counterfeits since the 3M masks are the low hanging fruit everyone is copying.
Capitalism contains the seeds of its own destruction. The relentless drive to shave fractions of a cent off costs in the short term creates the conditions you describe in the long term. In other words, it's like being penny wise and pound foolish on a global scale.
Fake News... that is not capitalism, that is a by product of decades and decades of tex based incentives that push people in to market index funds, mutual funds and other institutional investment where it is all managed funds
These managed funds are driven by quarterly earnings people expect to see...
This intern drives public companies to focus on quarterly earnings
Privately held companies almost never look at quarterly earnings, they look at a 5 year picture... It is mainly public companies majority owned institutional investors (401k's, pensions, etc) that focus on quarterly earnings
So once again what people blame capitalism for, is in reality government putting their hands on the scales and tipping the market in ways that have unintended consequences
> "service economy" was always a meme being sold to the gullible masses, manufacturing capacity is the only thing that matters for an economy.
Service economies work great in places where food and housing are free, so money is only needed to buy luxury goods. They only work poorly in the US because capitalists have leverage over service workers.
Service economy is kinda fun when you really think about it. Rather few sectors really make sense. Food, you can't make living just by selling food to each other, the ingredients and hardware has to come from somewhere. Hospitality quite similar. Banking? Banking and finance is just extracting money from others, and actual goods produced is quite minimal... Same can be said from quite many other field like law. Some of the work does increase productivity, but you have to have some base line of it to increase...
Because all of their competitors are also raising their prices. Why would companies reduce prices and profits when the demand exists even with higher costs?
If you look at prices for the stuff a lot of wafer capacity is going towards, particularly Flash and RAM, you'll see that the prices have gone up 20%+ over the last year.
I think people like to pick out TSMC because of the politics surrounding Taiwan and China, and maybe because a lot of companies use TSMC for their IC production and like to complain. People don't notice as much when other ICs like flash, ram, micros, etc are slowly increased in price.
People like to pick out TSMC because they're the winner of the foundry wars; GlobalFoundries bowed out of further process node development, I think two years ago, and Intel is trying and failing to equal TSMC's performance with their in-house fabs. GF and Samsung and other players are still around, but only TSMC offers the highest-resolution parts, so they're your only option if your part needs current fabrication technology, and they may be your cheapest option even if it doesn't.
There are three companies (TSMC, Samsung and Intel) that compete at the top end of chip fabrication in the whole world. One of them (Intel) didn't make chips for anyone but themselves until about 4 months ago. Preparing a semiconductor for a given companies fab process takes time. You cant just call them up and ask for however many units of your design. Its a months to year long process of your engineers working with their engineers to optimize your circuit design to work best with their methods.
Also there is only one supplier (ASML) in the entire world who makes a critical machine (extreme ultraviolet laser photolithography machine) needed for these cutting edge processes. It is unable to match the increased demand for their tools. And even if they were these super advanced might as well be magic machines have a lead time in years themselves as do the factories they go in to.
If you want to start marking modern semiconductors or increase your existing capacity to do so it isn't as simple as throwing a couple million at it and having it running in a few months. Its billions of dollars in a massive sterile factory using machines made by a handful of suppliers taking up acres of land and consuming impressive amounts of electricity and water to do so. Semiconductor fabs are massive sprawling industrial campuses https://images.anandtech.com/doci/13749/intel_kiryat_678_678...
Many missing chips for automotive or devices are not gpu or top of the line level but simple microcontrollers. Those can be done with 130nm to 40nm process for the most modern. 130 nm has been used for pIII or amd athlon64 and dates from 2001. Those need near UV, not extreme UV.
> One of them (Intel) didn't make chips for anyone but themselves until about 4 months ago.
actually iirc they've tried to get into the custom foundry business a couple times now (see: "Intel Custom Foundry" that got discontinued a couple years ago) but they never really attracted any major customers.
Part of this was due to the fact that Intel's design rules (the "rules" that govern chip design: how small wires and transistor features can be and so on) are completely different from every other foundry, so while porting from TSMC to Samsung or something like that certainly isn't easy, it's at least straightforward, whereas Intel's stack is basically totally different from everyone else's. One of the advantages Intel has is that the chip design and the process are tightly intertwined, if there is a problem with the chip design the designers can go down the hall and talk to the process people and see if they can come up with a solution at the process level. This obviously also became their downfall when 10nm slipped, and likely also it severely reduced their ability to sell their foundry services to other customers.
According to that article they were apparently also rather secretive about their process (seeing it as part of their "special sauce" I guess) and I've heard many customers doubted Intel's commitment to the program as well as their neutrality. And I think the final piece of the puzzle was probably price - I'm guessing Intel was asking a royal sum compared to GF or Samsung or TSMC. After all... they're intel, and their process is the best!
> One of the funniest stories I heard was about the first copy of the Intel 14nm design rules Atera got from Intel. They were heavily redacted, which is something I had never seen in the foundry business. After many delays Intel put their own implementation team on the first 14nm Altera tapeout and the result was a very competitive FPGA chip. If not for the continued delays, Xilinx would have been in serious trouble as the Intel 14nm FPGA, based on my experience with customers, beats the Xilinx 16nm in both density and performance.
They are now trying to get back into the custom foundry business of course. Different management, and a very different business situation for Intel - if they can get other customers to help finance their node development that's a big win for them at this point.
ASML is trying to increase the world's capacity by ramping up their production of EUV machines to 50 per year. [1] The ASML CEO stated it's probably not possible for the world to produce more than that. These machines are ~$150 million ea and building one is insanely complicated. [2]
Their max is about double that (other kinds of systems, and service.) That’s about a p/e of 60 which is too high but what this market has done to many stocks. There is some expectation of growth, too, of course. ASML will keep pushing their capabilities and if trends continue, they should eventually be selling individual machines worth an order of magnitude more.
What he's making is basically caveman technology compared to actual cutting-edge nodes.
No offense to his accomplishments, but the modern fabs like TSMC are so advanced you cannot just make the things they make without billions of capital investment. It just is not possible.
Unfortunately, many of the chips in short supply are on pretty small nodes, so you need extremely specialized equipment limited to a few of the big players to make them.
Certainly there's some stuff that can be made in larger nodes, but if you want something in 5nm you basically have one game in town.
Building a new factory is a 10+ year effort for a competitor. (They have to do the R&D to get on-par with performance, and physically build the factory).
The shortage won't last that long.
But in the meantime, anyone who owns a suitable factory can pretty much set the prices as high as they like. I'm really surprised the price hike is only 20%, rather than 10% per month for the next 3 years.
Which government? Taiwan, where it matters? This is a major part of their economy; they all win big. I bet they'd sooner surrender to China.
Some other government? This wouldn't have much impact on chips, except to make it a lot less likely anyone makes chips in that country ever again — which isn't to say no nation would do it; economic history is full of this sort of thing, particularly as nations start going the way of Venezuela.
That just disrupts the profit incentive of competitors.
The entity that manufactures the product is compensated for it, meaning the money isn't being siphoned off by a random third party. Price controls grant a wind fall to consumers who are in no position to build another semiconductor plant. It's basically a wealth transfer away from the semiconductor industry into the consuming industries. If you want semiconductor prices to drop then either grant a subsidy or let the government build a plant that produces way more than the market really needs.
A single shortage plus price increase might not be able to pay for a whole plant, but any plans to expand capacity will be shifted forward rather than be delayed. If the profit motive exists the construction of the new plant will happen one year sooner and it will be bigger in anticipation of future shortages.
Probably the limiting factor is one needs to make sure the customers of chips stay viable. I mean, one hopes the shortage will eventually be over and foundries will want to make sure they didn't gut their own market.
From my experience no one cares if they are as long as they more or less work.
Source: I'm a software engineer who has had to fix so many issues due to unannounced component changes from suppliers.
"Oh you just replaced the specific USB chip with something different that has a whole lot of errata (Bugs in chip logic that you have to work around in software) and didn't tell us? GREAT! THANKS!"
Oh man. I just did a board bring-up recently and didn't notice that the flash chip I was having trouble with was a previous version than the one we specified. I only just barely managed to get those boards working, and best as I can tell, it was the conductivity of the particular flux I was using to reflow the part that was nudging an NC connection to actually use it's internal pull-up. When the factory couldn't follow my steps, they finally realized that if they were going to substitute that chip, they'd need to add a 10k pullup resistor on a pin that had a use on that revision but not on the one we wanted. What's worse, is even if I'd noticed that the chip was older, looking at the datasheet for it that I had access to, nowhere did it mention a need for a pullup on that pin. Only the factory datasheet mentioned issues with the internal pullup and suggested using an external one.
Supply shortage is another way of saying demand excess. I don’t see much evidence that TSMC ever really slowed production, at least for their most advanced processes that are in high demand.
This is yet another way we’re all forced to pay the price for cryptocurrency. As long as GPUs are profitable for mining, they will be purchased and put to work. The saddest part is that each additional GPU added to the mining networks only makes the overall system less efficient. Mining difficulty is automatically adjusted upward to cancel out the additional computing power. The networks processed just as many transactions last year as they do now, but countless new GPUs have been added to the mining network due to the incentive structure of mining. It’s a race to the bottom, consuming ever more GPUs and energy until equilibrium is reached.
It’s frustrating to watch so much of the output of our most advanced chip manufacturing be diverted straight into making arbitrary proof-of-work systems less efficient so someone can capture incrementally more coins. And now we must all pay higher prices for everything.
Estimates on Bitcoin energy usage put it at over 1TW now. I believe it was comparable to the energy consumption of Argentina.
Advocates will defend this by saying most energy usage is renewable. This conjures the image of someone with a bunch of solar panels but the reality is that it's primarily hydro power because that's the cheapest. Thing is, in regions with cheap hydro power, the miners can use so much power they end up making power more expensive for everyone.
I'd be more OK with this if crypto in fact solved a problem for most people. We should start by stopping calling them "currencies". They're not. They're assets. They lack all the useful properties of a currency (eg being massively deflationary)
The only thing cryptos really do is allow a temporary medium of exchange for traditional currencies. Some uses of this are entirely legitimate (eg escaping capital controls in certain currencies). Some are not (ie all the illegal usages).
I really wonder what happens to Bitcoin (or any PoW coin) when it no longer becomes economical to mine new coins. Does all the computing power just move on to the next coin? If so, won't this make the network vulnerable to attack? What is the incentive for people to contribute computing power without the prospect of economical new coins?
> Advocates will defend this by saying most energy usage is renewable.
Even if true (iirc there is a lot of doubt whether it is, see Iran), it's still moronic, because consuming renewable energy for shits and giggles obviously means that other energy demands are being met with your usual energy mix instead of that renewable energy that went into shits and giggles as assumed. In terms of net emissions it simply doesn't really matter whether you A) use non-renewable energy for BS or B) use renewable energy for BS, displacing other energy users to non-renewables.
The simple fact of the matter is that there is no free lunch and there is NO ecological way of randomly wasting energy.
Now, the claim could still theoretically be true, but I really want to see some hard numbers with sources from the next person trying to defend that claim.
Subsidized energy can be cheap or even free, and it's hard to compete with free. For-profit Bitcoin miners are in an intense competition to find the cheapest energy. Anyone who can steal their energy or get someone else to pay for it via subsidies will win there, but only until the free lunch ends. Energy free lunches are ending all over the world now; nobody wants to build a coal plant in order to give away the energy to bitcoin miners so they can retire in Switzerland.
And that's why the Chinese authorities kicked out all the bitcoin miners a couple months after that, dropping the bitcoin hashrate by more than half. It's on its way back up, but it hasn't fully recovered.
By contrast, especially in tropical regions, you can build a solar or wind power plant to mine bitcoin profitably without finding someone else to fob the bill off on. In much of the world, solar energy is now less than half the cost of fossil-fuel energy, though not in cloudy polar regions like Germany, the Netherlands, or the UK. There's a new record-low PPA price every few months, and it's always solar; https://www.pv-magazine.com/2021/04/08/saudi-arabias-second-... this April seems to be the latest, at US$0.0104/kWh, which is about a quarter of what coal energy usually costs.
That's why the Navajo Generating Station has been demolished, Peabody Coal has gone bankrupt and may do so again, and for the first time in human history, outside of PRC, more coal generating capacity was shut down than was built last year. This year that'll probably be true worldwide. Coal (and nuclear, and oil) just can't compete economically with solar anymore, even in temperate regions. And, consequently, neither can Bitcoin miners who are paying the cost of using coal. Fortunately for them, shipping their server racks to a more profitable region is a lot cheaper than doing the same with a steel mill.
Hopefully this gives you a clearer picture of the whole political-economic environment in which Bitcoin mining is happening than just a few hard numbers, which are of course fundamental.
Proof of work cryptos are a pox on society. That said they are some of the most location neutral energy consumers. This allows mining farms to be located adjacent to hydro dams or wind farms to get the cheapest energy and use energy that would otherwise be lost in transmission. 1TW is around 20% of global energy production. This is enough of a factor to impact manufacturing of renewable energy equipment. It allows them to achieve economies of scale faster and accelerate the transition to cleaner energy that we need.
It's estimated all of the U.S. is ~17% of global energy. [1]
"According to the Cambridge Center for Alternative Finance (CCAF), Bitcoin currently consumes around 110 Terawatt Hours per year — 0.55% of global electricity production, or roughly equivalent to the annual energy draw of small countries like Malaysia or Sweden." [2]
The incentives to use renewables are there... where the costs of fossil sources are high, and where the energy consuming process will be reliably profitable long enough to pay for all those renewable energy sources.
That last bit makes me doubt this would push renewables - who trusts that mining will remain profitable for years or even a decade? In other words, it may well incentivize energy sources with low down-payment costs.
Also: all those GPUs are expensive; is a miner really going to slow down mining significantly just because it's dark or no longer windy? Or will they prefer constant power, given the fact that GPU's will become obsolete much more quickly than the power source, and thus there's pressure to recoup that investment as quickly as possible, I'd guess... thus incentivizing baseload power, not unreliable but cheap power - conceivably, anyhow?
The high bit here is that coin mining is not 1 TW; it's 0.012 TW, so the overall effect is almost insignificant. But it's still interesting to know which direction the incentives run.
The cost of fossil-fuel electricity generation is about half capex and half opex (including fuel), and in much of the world, solar is now cheaper than either fossil-fuel capex or fossil-fuel opex. A large part of capex is machinery, much of which is so costly that shipping costs are insignificant, so the costs are the same anywhere in the world.
Bitcoin isn't mined with GPUs, but your comments about intermittency and capital costs apply a fortiori to ASIC mining.
>This allows mining farms to be located adjacent to hydro dams or wind farms to get the cheapest energy and use energy that would otherwise be lost in transmission.
Sure, but there doesn't seem to be to be any indication that this is what actually happens. Is there any evidence for that?
If only we could send a self-replicating machine that builds raspberry pis and solar panels to the moon with an uplink.
Then the difficulty would jump and we'd need to send one to Jupiter.
Then to mercury to set up the dyson swarm.
Suddenly we are a Kardashev civ... but it's all for bitcoin mining for rich billionaire paranoia of centralized currency control by the US government, and the actual bitcoin production rate is constant since the difficulty rate on the coins exploded from galaxy-spanning miner fleets.
As explained above, the real number is closer to 0.012 TW, so while the impact on the renewable transition is probably positive, it's very small, more like 0.1% of global energy production.
It's disturbing that nobody in this whole thread thought to fact-check those implausibly large numbers; it indicates that the whole thread is dominated by herd instinct, not rational thought.
I think Bitcoin is ridiculous and wasteful, but I'm also conflicted because I think similarly about many physical products: jewelry, mansions, oversized personal vehicles. There's a lot of embedded energy in all of those goods. Who am I to judge which ones are worthwhile. It all makes me dearly wish there was a global tax on taking the resources out of the ground.
I'm curious why we don't find a similar backlash to people who put their energy into a hot tub or to cool a 5x sized home.. as these cases don't even purportedly support a valuable economic mechanism.
For instance. I don't cool my home with AC. does it bother you that I mine bitcoin overnight when the grid has surplus energy?
People do complain about excessive comforts too. It's one of the main complains those less well off make about those more well off.
That said, there's a couple of differences between hot tubs and AC, and crypto mining:
- Creature comforts don't consume the energy equivalent of a small developed country.
- In what they do consume, most of the use is around basic human needs; excessive use happens rarely.
- Most importantly, for almost[0] everything other than crypto, whether it's hot tubs or ACs or fiat banking system, energy use is considered waste to be minimized. Simply because energy is not free, the market is incentivized to reduce marginal energy consumption involved across the board[1]. In contrast, PoW considers waste a feature.
Cryptocurrencies are their own special thing - the perfect machine for consuming any surplus energy available. That's why they deserve their own special kind of backlash.
--
[0] - I can't really think of any thing other than PoW cryptocurrencies, where energy waste would be beneficial to the thing by design.
[1] - That, of course, doesn't mean we're using less energy overall - because both population and economic growth outpace efficiency improvements. But on the margin, it holds true.
- Creature comforts don't consume the energy equivalent of a small developed country.
you are incorrect here. 10% of global energy (not just electricity) is used for cooling.
I live in the deep south with no AC. "basic human needs; excessive use happens rarely." is extremely generous.
What you are describing is what makes you comfortable. I have no problem with that. be comfortable. Just consider that what you are describing is completely a matter of perspective and not some universal truth about use of energy. Just like you are not universally correct that every watt spent on crypto is "wasted". It's a perspective.... and my point here is to remind you to keep it in perspective of other trivial preferences people have, rather than turning it into an ideological war of good and evil.
To add, this isn’t really “surplus” energy that’s being consumed.
There are fossil burning power plants running only to produce PoW, and unless mining operations shut down to allow a generator to operate at minimum load, they are forcing shared generators to run and dams to drain.
you know that generator fuel consumption and dam water consumption is dependent on the load on the generator, right?
it's not like a generator burns 2gal/hr no matter the load. This is easily observable even (perhaps especially) on home generators - if you run it idle it might run for 12 hours and if you load it to maximum capacity it might be out of fuel in 2 hours. look at a product page for a generator and will probably be specified at a particular load level (50% load for example).
also if you keep pushing water through a dam at maximum rate even when the dam is under a light load, you'll overspeed the generator and blow it up. You have to have control gates to control the water flow.
basically, "how difficult" it is to spin the generator is determined by the electrical load (coupled through the magnetic field of the motor) and obviously you need more physical power to spin against a stronger field, which translates into more power/fuel.
A lot of people are in the habit of complaining about what other people do that might possibly but not likely indirectly affect them somehow. Of course the powers that be are fine with this (and probably promote it) because as long as the populace is pointing fingers at each other's lifestyle, they are ignoring the massive use of fossil energy (among other things) used to make a few rich people richer.
"Penny wise and pound foolsh."
Another example of this is plastic bottles. See it's your responsibility to recycle, not the company's responsibility to not use plastic. People nanny each other's recycling habits but don't say shit about Coca-Cola bottlers. Bottler's are too busy laughing their ass off on the way to the bank. Recycle, reduce, reuse.
In the case of cryptocurrency, there is a clear better alternative that is Proof of Stake. It's caused by techological, economical and educational inertia. Cryptocurrency is driven by community consensus, and that's what we're doing here... discussing an outdated consensus (PoW supremacy) in the hope of changing it.
Proof of stake changes the game and raises the bar for entry. It essentially creates a "rich get richer" scenario where the only way to increase the amount of coin you have is to either already have enough to participate in proof of stake, or to buy the coin from someone else. In that light, I wouldn't necessarily say that proof of stake is clearly better than proof of work.
As an example, the proposed proof of stake mechanism for Ethereum requires 32 ETH, which is around $100k USD today. So established parties with capital will be able to stake, but your average person will not. I'm not sure I like that aspect of it.
It seems there will be pool staking in ETH, although I agree ideally staking threshold should be low.
Consider the case with Proof of Work (ASIC-based): you need to purchase specialized hardware for several $k (apparently at least $2k), spend more on specialized infrastructure (power supply, cooling, servers) and live somewhere with almost free energy. Otherwise it's just not economical to mine. This leaves out the vast majority of people as well, and incentivizes (1) massive economies of scale of farms, (2) shady or corrupt deals for free energy, through local government or simply illegal connections to the network.
Ridding the world of centrally planned currency isn’t shits and giggles. We’re in the midst of the largest wealth transfer in history. Trillions are being given to the wealthy while the poor see their purchasing power wither away. I’m not certain we’ll successfully escape their depraved clutches, but wasting a little electricity is worth the chance.
I still fail to see how it addresses wealth distribution issues.
Crypto currencies are just an asset as any other financial asset.
Yes it's decentralised but the only thing it seems to have proved so far is that centralisation might not be an issue in the first place.
Since crypto currencies are more of a financial asset than a currency, it's also highly dependent on actual currencies in the end.
Because the powerful people with more crypto than you can't magically issue infinite more crypto and give it to themselves, devaluing the crypto in your wallet.
Classification as an asset is part of the state fighting against crypto currency. It's a non sequitor. If the dollar was deflationary more people would be holding onto it as an asset. But the dollar is intentionally manipulated/inflated so that it's a liability to hold it, therefore people trade it for other asset classes.
The difference is that the powerful people with a lot of crypto can't unilaterally decide to issue themselves trillions more bitcoins which devalues the few coins held by the poor. Yet this is exactly what happens with the dollar.
The energy use is of captured energy that would be lost in transport to other energy users, because it uses just as much energy or more to transport it.
This means that “renewable energy more than the output of Argentina was always being wasted for over a century every day because it was uneconomical to transport it”
Crypto mining has been unique in that it only needs power and housing to stay dry, low infrastructure, and doesn't need very good internet, compared to other large computational projects, or any other sector.
Okay, so we got the “ah it uses a lot of energy!” argument over to “ah okay its a relatively green sector compared to literally everything else but I don't respect the use in favor of literally everything else and this must be taking away from that” which is also not true so I’m curious how the goal post moves after this - assuming you are able to corroborate what I’ve told you at face value.
It would be more productive to advocate for specific sites to not raise prices for others if they did, or specific sites to use cleaner energy sources. I can understand how that could cause some consternation if your whole cause is not liking crypto mining at all. But it is important to remain vigilante on how these mining operations run, as it is possible for nation states to operate these uneconomically and actually do things that you imagine is happening now.
(Renewable and captured are not synonymous but usually its the same)
If your first paragraph is about rejected energy, then there's not a lot to say here except that rejected energy doesn't mean what you think it does.
If not, then I don't know what it is supposed to be about. No one builds power plants in places where there is no demand for power. Electric power can be carried over pretty large distances though with very low losses, so having say a few hundred km between a hydro-dam and a large consumer is not a major problem.
According to your post the best result would be that Bitcoin farms are powered by renewable energy plants build in places where nobody would have any use for that power anyway, because they're too remote to connect to any consumer. Even if this happens (and from what we're able to tell, it is not, instead we see coal and natural gas plants being ramped up to power bitcoin farms), it's still a net negative because building renewable energy power plants (as well as bitcoin farms) causes a significant amount of emissions. In normal use of e.g. a solar panel, this makes sense because it displaces non-renewable energy sources with higher emissions, so net emissions are reduced. In your scenario, there's a bunch of solar farms in the middle of the Saharan desert mining bitcoins, which - at this point - yes would not be able to displace non-renewable energy, but still represents a net-negative regardless because of the effort required to put all that shit there in the first place.
The simple fact of the matter is that there is no free lunch and there is NO ecological way of randomly wasting energy.
> (and from what we're able to tell, it is not, instead we see coal and natural gas plants being ramped up to power bitcoin farms)
like I said, advocate against those specific mining sites. those made headlines in your world, while others did not.
A lot of captured energy mining occurs at remote flare gas sites, where those sites are in the business of pumping oil and flaring off byproducts and not in the business of moving power a few hundred km. These are highly polluting places, and crypto mining is ironically (to you) their sustainability solution, as they provide a use for the energy that would have been flared into the atmosphere as hydrocarbons instead. As these are also fracking sites, the machines can use the polluted water for cooling. Those kinds of sites are only captured and not necessarily renewable, an important distinction because your example was limited to imagining uneconomical solar. There is much more energy like this able to be tapped while reducing hydrocarbons in the atmosphere, and if it was economical for other use cases that you respect more to tap into that energy, you and others had several decades to do so.
It's cheaper for them to just flare the gas. The only thing that has made economic sense, as an alternative to flaring it, is to mine Bitcoin. I understand the energy criticism, but flare gas is an example where Bitcoin mining actually makes something less polluting.
- It's approximately a 180° pivot to go from "unused renewable energy" to "flare gas for green bitcoins"
- It doesn't reduce the emissions. Whether you flare it or burn it in a gas turbine / engine driving a generator to mine bitcoin, the CO2 released is exactly the same, as you are burning ~100 % of the hydrocarbons either way. The difference is that one generates bitcoins, the other does not. This is only valuable to bitcoin stakeholders, and as has been pointed out up and down all these threads this is either of approximately no value or of negative value to society at large. And overall, creating the equipment to mine bitcoins with flare gas, and deploying that to remote oil fields w/ no use for gas, creates a bunch of emissions that simply don't exist if you just flare the gas.
- There are non-emissive alternatives, e.g. gas re-injection. Note that the great majority of unwanted gas is already dealt with this way. This means that mining bitcoins using flare gas has a very good likelihood of increasing the share of unwanted gas that is effectively flared, again, increasing emissions.
Reducing the amount of flared gas (and putting it through a turbine to mine bitcoin is the same as flaring it as far as emissions go) is in fact an UN IEA goal.
The simple fact of the matter is that there is no free lunch and there is NO ecological way of randomly wasting energy.
> - It's approximately a 180° pivot to go from "unused renewable energy" to "flare gas for green bitcoins"
I wrote earlier to avoid this specific derailing. Renewable and captured are not synonymous but usually the same. This is one of the cases where it is not the same. To my knowledge, the only other captured sources are at hydroelectric plants, while some parties are interested in geothermal as well.
Noteworthy is that “captured” seems to be also called “rejected” energy in some circles.
> And overall, creating the equipment to mine bitcoins with flare gas, and deploying that to remote oil fields w/ no use for gas, creates a bunch of emissions that simply don't exist if you just flare the gas.
Once. I’m not sure how a one time event is remotely comparable to ongoing issues. Please elaborate?
> It doesn't reduce the emissions. Whether you flare it or burn it in a gas turbine / engine driving a generator to mine bitcoin, the CO2 released is exactly the same, as you are burning ~100 % of the hydrocarbons either way. The difference is that one generates bitcoins, the other does not.
Hmm. Is there anything we can do about that then? Like does that format allow for further processing towards sustainability? The generating of bitcoin making it economical to do more with the CO2 that the oil guys just threw their hands up on decades ago? If this issue has nothing to do with bitcoin mining and the alternative is just business as usual with it going directly to the atmosphere like it already was, then why point fingers at the bitcoin mining?
Thanks for chiming in! Nice to see a new account understanding this issue and joining the fray.
It has been fascinating to me to watch the disbelief evolve. One aspect of this is that negative reporting has been basically allowed and encouraged by miners because they didn't want competition on their cheaper renewable/captured energy sources.
So, for years, many investment groups would discourage their renewable or captured energy activity to be citable source of what is happening with bitcoin mining, because they didn't want others with deeper connections to make partnerships with sites faster.
Its really funny, to me, watching people have to redo their understanding but trying to make that understanding conform to the negative energy impact idea that they were intentionally led on with.
> OK, but those exact same places could power servers instead of miners, taking a load off of wherever the servers are currently hosted.
They don't have good internet. Which I wrote earlier. Crypto mining don't need good internet, it barely needs any good internet speed and barely needs decent latency - by design (blockchains allow anything from 2 seconds - 2000 milliseconds - to 10 minutes - 600000 milliseconds - for the majority of miners on the whole planet to reach the same state together, under the assumption that some nodes will have slow data and low latency. So a 500-900ms latency connection in the middle of nowhere is quite fine, not optimal but good enough for consistent average earnings). Other data heavy computational uses require the opposite of that, these remote sites are not going to get the plumbing for good internet or prioritize it. The main issue is that they won't prioritize it because they aren't interested in it, they are technophobic oil guys that are slowly being convinced by crypto investors to solve their pollution problem. The crypto people don't need anything except a little space.
The issue with imagining that everyone else could do it, is that everyone else doesn't need to do it. A data center can set up in the middle of the desert along any fiber line like they already do. Go ahead and tell them "hey instead of using this readily available property can you go set up at the oil refinery after making an unnecessary partnership with that unrelated business because I don't like that crypto people are doing it, you also need to invest in laying some pipes out there" if you really think that is practical or necessary, be my guest.
Yeah this shouldn’t have been flagged, this part of hackernews is very broken for them to inadvertently allow censorship like this
The stackexchange approach would be better here, where certain power users can vote on flagging and you can see who voted on flagging, other users can vote against flagging and even override it
Mining is definitely renewable UNfriendly. Hardware is too expensive to turn it down at night, or when winds are weak. It's 24/7 constant demand, the worst possible for solar/wind.
It's indeed location independent, but that's more compatible with previously uneconomical fossil fuels (eg. methane in locations where transferring it out isn't economical).
People need to remember that using up the available renewable energy on crypto just means there's less available for other consumers to use. The idea is to reduce electrical demand which makes it easier to shift towards 100% renewable. Adding all that electrical demand just keeps around non-renewable energy longer since few countries are able to go 100% renewable on their current demand.
I doesn't, at all. Energy production and energy distribution are entirely different things. There is enough unused (i.e. wasted) hydro power in Quebec to power the entire Bitcoin network. If energy is produced too far from where it is demanded such that transmission is prohibitively expensive, there is no downside to having highly mobile miners to make use of it. In fact the extra revenue generated can be used to subsidise the rest of the grid.
Here's a writeup on how that's a misleading statement, and how the best use of that hydro power for offsetting carbon would be for it to be used to offset neighboring province's much worse coal-based power.
Transaction fees will always make it economical to mine. They currently make up a small fraction of the block reward, but within a decade or two (5 block subsidy halvings), we can expect them to dominate.
Btw, not all cryptocurrencies are deflationary [1].
It won't if people realize that BTC is not a useful product, and the price falls to $1k or $100. At that point no one except someone stealing electricity can afford to mine, so the entire monetary system collapses and becomes worth the same as US Civil War dollars (can maybe buy a loaf of bread with a wheelbarrow full?)
Given that bitcoin has been around for 2008, at what point will you personally concede that it's not likely to happen? Why would it happen now if it didn't in the first 13 years?
Madoff's scam lasted for 17 years and only collapsed because of sharp economic downturn. It could have lasted for much longer if people suddenly didn't need the cash.
Bitcoin is often described as distributed Ponzi. There is little value (skipping over regulations is about the only thing it's useful for) and a lot artificial pumping. It will collapse when people need money for something else so either economic downturn (exactly the thing crypto believers think they are hedging against) or when it slowly goes out of fashion (probably either due to regulation catching up or something much better coming about). Till then the show goes on. Even if there is just 10% of real dollars behind those valuations it will not matter until there is a bit of a run to cash out.
Real estate does something. If everything else goes to crap, I have a place to build my home. My home has a pretty high base value. If full banking collapse and we go back to trading, I would trade my house for a LOT of apples.
What does BTC do for me? It's a number on a screen. Much like a NFT.. it has no actual value on it's own. In a zombie apoc, no one is trading you food for BTC.
Stocks are not Ponzi wtf? Even today you can get something like Apple making 3% per year while also having big chance of making more in the future. It has nothing to do with a Ponzi scheme at all while artificially propped up producing absolutely nothing crypto currencies have a lot of similarities.
The argument that scarcity is a fundamental value proposition for an asset is just nonsense. I can make my own crypto even more limited than BTC overnight and it will worth about 0. It's not scarcity that makes stuff valuable. In case of BTC 3 things make it valuable: adoption, ability to make transaction while skipping over regulations (which requires adoption) and artificial propping by scams like Tether.
Stocks are a ponzi by definition.
All growth in stock prices are dependent on extracting money from future investors. It gets a little more complex with corporate stock buybacks, but ignoring this, the price of a stock tracks a companies success, but is ultimately a product of the market buy/sells. Without injection of funds by the corporation, its a Ponzi like anything else. Even more so for stock that does not issue dividends.
Same goes for housing, or anything else with scarcity.
Wall street has done a great disservice making you believe the price of a stock accurately reflects the value added by the corporation.
I agree that adoption (network effect, Metcalfs Law), censorship resistance and other economic constructs like stablecoins all have an influence in value. Folks have been saying "I can make my own, scarcity is bunk" for a long time now. 10,000 other clones of bitcoin exist today that are worth nothin'. Maybe there is something about big b Bitcoin and the cryptoeconomic model that Satoshi chose.
The idea of a stock is that it is worth the present value of all future dividends. Stocks that do not issue dividends and are expected never to issue dividends are indeed a Ponzi. I think those are by far the minority of stocks.
The idea that stocks that never issue dividends are somehow only valuable because other people think they are valuable (let alone calling it a Ponzi) is completely misguided. I mean if a company builds a pile of cash and never distributes it to shareholders it's still very valuable to have a say about how that pile of cash should be spent. If a company ammased a billion of cash there will be many takers to buy it for less than a billion just so they can direct that billion somewhere.
Of course there is also an option of closing shop and dividing the assets to shareholders or myriad other ways of making the assets valuable without paying dividends.
Those kinds of basic misconceptions about financial markets is what crypto evangelist prey on. It sounds pretty convincing to the ignorant but is just nonsense under scrutiny.
Man, companies make money or have good prospects of making money in the future. You own part of the profits. If the company pays dividends doesn't matter at all. It's econ 101 or not even that.
I think it's a bit disingenuous to use 2008 as the point of reference. Bitcoin didn't really reach the point of public consciousness before the 2017 boom (which was largely attributable to price manipulation. Prior to that the main use-case was buying drugs online, and prior to that it was a niche hobby for libertarian futurists.
Given the increasing prevalence of ransomeware attacks, and the fact that they have touched higher and higher profile targets, it's not impossible to imagine the regulatory hammer coming down on cryptocurrency at some point in the near-ish future.
It's going to be a flash and a poof though. I don't think it's going to go down 10% a year for 10 years or something. I do think it may drop 50-70% in one day/week, which makes miners shut down, which causes a full collapse.
That's only true if people are actually willing to pay that much in fees though. If the amount from fees doesn't keep up then eventually some miners will have to stop mining.
That's certainly not a good example. The price has grown much faster than inflation in the western world, while the issuance of new currency has trended down, not only as a percentage of total money supply, but in nominal terms:
https://www.wisdomtree.com/-/media/us-media-files/blog/blog-...
While in theory the community can elect to inflate the money supply, in practice the people who get to vote on the issue are holders of currency, who have strong incentives to restrict the supply and deflate the currency. If an inflationary EIP is proposed, these people will simply ignore it and stick to the deflationary chain.
There is always confusion when discussing deflation and inflation because there are multiple definitions. There is a inflation/deflation which refers to the size of supply and then price inflation which refers to the cost in some other fiat currency. The former is simple and clear, the latter depends on other variables.
You are ignoring the large amounts of coin that are taken out of circulation by software and hardware bugs, owner death or amnesia, intentional destruction by transfer to unspendable addresses. For bitcoin, by credible estimates we are already talking about millions of coins.
No, https://cbeci.org/ says Bitcoin uses 0.01176 TW, not 1TW. You can take a deep breath of relief! The real energy use of Bitcoin is only 1.2% of what you thought it was!
So, if you thought Bitcoin energy usage was a pretty bad problem but not world-endingly dire, you should now think it's insignificant. If that isn't what you think, it probably means that your opinions are not based on careful cost-benefit analysis, just partisanship, and should therefore be disregarded.
Energy consumption is measured in terawatt hours, not TW, and page you linked to says annualised consumption is 90TWh , which is on par with the netherlands annual consumption. This is hardly partisanship, its an enormous amout of power.
> Energy consumption is measured in terawatt hours, not TW, and page you linked to says annualised consumption is 90TWh , which is on par with the netherlands annual consumption. This is hardly partisanship, its an enormous amout of power.
You can measure energy consumption over a given time period in terawatt hours; for example, consuming a terawatt for one hour is consuming one terawatt hour, while consuming a terawatt for two hours, or consuming two terawatts for one hour, is consuming two terawatt hours. The standard term for "energy consumption over a given time period" is "power", and the standard SI unit for power is the watt. A terawatt hour per year is precisely and identically some number of watts; I calculate it as 114,079,550 watts, or more conveniently 114.1 megawatts, because an hour is about 0.00014407955 years.
If you want to measure power consumption in ad-hoc units of 114.1 megawatts or poundal-fathoms per Planck time, I can't stop you, and you can even do correct calculations with them, but, myself, I prefer to use SI units, because they greatly facilitate mental arithmetic and comparisons.
(This has been a brief introduction to the basics of the Systeme Internationale of units.)
If you scroll up about 512 pixels from where you were looking on that page, which currently says annualized consumption is probably about 93.01 TWh/year, you'll see that they also provide the same number in a different form, as 11.76 GW. (This is a slight arithmetic error; 11.76 GW actually adds up to 93.01 TWh every 330 days. Perhaps the 11.76 GW is a 7-day moving average and the 93.01 TWh/year is instantaneous power consumption, or vice versa.)
Whether 11.76 GW is "an enormous amount of power" or not depends on what you're comparing it to. Total world marketed energy consumption is about 18-20 TW, so compared to world marketed energy consumption, it's not an enormous amount of power. (A terawatt is 1000 gigawatts.) Terrestrial insolation is about 130 petawatts (a petawatt is 1000 TW); about 0.3% of that falls on arable land, so agriculture uses about 350 TW of power, almost 20 times world marketed energy consumption.
Unsurprisingly, you're also mistaken about the Netherlands' power consumption, perhaps because you don't know about non-electrical forms of energy such as kinetic energy, chemical energy, and (very ironically in this context) solar energy; even their electrical energy consumption alone is 119 TWh/year, which is 13.6 GW, but most of their energy consumption is not electrical; their total energy consumption is more than 100 GW.
So, Bitcoin provides a secure digital currency to 7.8 billion people for, probably, significantly less energy than the Netherlands uses to provide electricity to 0.018 billion people. (A million is 0.001 billion.)
Is it partisanship to call that "an enormous amount of power"? Maybe, maybe not, it's a valid subjective perspective. But it's clearly partisanship to not care whether it's 12 GW or 1000 GW.
I agree that the externalities are still overwhelmingly negative, but one part you're missing is that cryptocurrencies act as a global reserve asset. Crypto is still tiny compared to other financial markets, but the existence of cryptocurrencies has probably reduced the volatility of other assets somewhat.
In the future, this effect will likely be more pronounced. It's an asset class that has some degree of independence from other asset classes, so it decreases the overall risk of the financial system.
The only good quality of Bitcoin is that it is completely decoupled from the economy. A bursting Bitcoin bubble doesn't take the rest of the economy with it. However, that only holds if countries don't use it as legal tender.
This isn't true. There are Bitcoin collateralized loans already, and there are Bitcoin backed ETFs. Companies have issued bonds underwritten with valuations that include BTC in the company treasure.
Bitcoin is small compared to the rest of the economy, and it's less integrated than many other asset classes, but it is very much coupled to other financial markets.
The problem crypto solves is you can memorize a seed phrase and the government has nothing to seize, no bank account to freeze, no physical object to take. It's value that resides completely in your mind, but is recognized by others willing to pay a market rate for it.
You're joking yourself if you think government seizures are just; in the US asset forfeiture does not require even a criminal conviction. This is a problem for all people. Crypto solves a problem for all people.
-money transfer and storage for the structurally unbanked.
Anonymous crypto wallets cannot be banned and to "taint" any bitcoin that has passed through anonymous wallets would mean every single bitcoin would be tainted. It would be like trying to ban euros used for drug trade; if it were done all euros would seize to circulate.
Don't you think it could be aggressively regulated against? I doubt it could be completely stopped, but demand could be sharply reduced if using and mining cryptocurrency carried sharp legal penalties.
It could, but governments have a hard enough time sharply reducing drug use, and that has something tangible to search for and find. Trying to find some bits in the ether or even worse guess whether someone is memorizing a seed phrase is nigh impossible to prove or extinguish.
Mining will be carried out wherever someone could get away with it, which is basically any western country where it is legal and any 3rd world country with internet access and power. It would be like trying to stomp out coronavirus in the entire world -- when you apply the "vaccine" somewhere, the reservoir just lives somewhere else.
The issue is while the legal demand will drop, the illegal demand will remain, which is the only one you're trying to stomp out. As a deflationary currency much value will be lost, but it will retain pricely enough market cap to sustain the black market.
If mining goes down as hardware capabilities go up, it’s only a matter of time before somebody fires up a mining operation for a few days to outcompete the rest.
It doesn't right up until the moment that somebody realizes the reward of profit is worth the risk of operating a mining operation capable of pulling off a 51% attack.
I said Bitcoin's survival was not predicated on centralized mining facility. You seem to be agreeing with me, because your scenario is one of a 51% attack predicated on a large centralized mining facility. You are overlooking the number of actors capable of pulling off illegal mining, and the difficulty in coordinating all of them.
You are acknowledging even under clamp downs, mining will continue. If you think about 30 seconds further, you'll realize it will be widespread and it is likely not just one clever 51% attack hopeful will come up with the bright idea of running an underground mining operation. Prisoner's dillema keeps all the decentralized nodes from colluding (risk losing your mining reward against the hope everyone will stick to the deal and hash a dirty chain). Prisoner's dillema acts the same way against more centralized groups so long as no one of them controls a large portion of the available hashrate.
Basically your statement is "people will pull off scaled mining operations even if illegal, and bitcoins demise is predicated on centralized mining operations" and I agree with you 100%.
If your argument is that Bitcoin is susceptible to a 51% attack, the answer is you are correct and you always have been. It is built into the protocol.
Bitcoin's survival is predicated upon the distributed, cooperating members of the network being willing and able to spend more cycles than any one bad actor.
Prohibition of mining centralized mining tautologically reduces the number of cycles the network is able to cooperatively dedicate toward mining. Whatever the hash rate otherwise would be without a prohibition, once you prohibit and drastically reduce the contribution of centralized mining, it would be drastically less.
But the ability for someone to put together a centralized mining facility is still there, it just carries additional risk. If nobody else is performing centralized mining, the incentives for someone to quietly put together an outfit of their own for the sake of attempting a 51% attack grows. This might be outright illegal, it might be done by a government, or a myriad other possibilities. Either way, it's an order of magnitude easier in that future than it is today for someone to gather the necessary hardware to mount such an attack and turn it on all at once.
> Basically your statement is "people will pull off scaled mining operations even if illegal, and bitcoins demise is predicated on centralized mining operations" and I agree with you 100%.
My statement is "people can pull off scaled-down mining operations even if illegal, but this reduces the effective hashrate; further, other people can pull off scaled-up mining operations in the short term".
>Bitcoin's survival is predicated upon the distributed, cooperating members of the network being willing and able to spend more cycles than any one bad actor.
Which is an entirely different statement than saying bitcoin's survival is predicated on centralized mining. It's not.
If centralized mining operations can continue, centralized mining operations could have a 51% attack.
Thank you captain obvious, ultimate king of semantics.
Your entire argument is one of bitcoin's demise via 51% attack being predicated on centralized mining. I don't think you understand that claiming a 51% attack is predicated on scaled-up operations is an entirely different statement than arguing against bitcoin's survival not being predicated on centralized mining.
You imagine a scenario of majority hashrate bad actors, all cooperating, can pull off centralized mining, and that their hashrate exceeds the entire hashrate of the decentralized less upscaled nodes as well as the more centralized good actors. That is a possible, but unlikely scenario that has always existed since the beginning of bitcoin for the reasons I previously stated before. It simply hasn't happened, there is currently a roughly $1T bet that it won't happen in the near term, and there is a valid argument to be made that when centralized mining becomes more restricted that the likelihood of successful 51% attack goes down and not up.
>people can pull off scaled-down mining operations even if illegal, but this reduces the effective hashrate
Probably, but not necessarily so. If I shutdown my up-scaled operation and send my miners to 100 of my friends to use discretely in their homes with similar uptime, tell me what difference does it make to the total hashrate? The only real effect is it makes it even harder to coordinate a 51% attack.
>My statement is "people can pull off scaled-down mining operations even if illegal, but this reduces the effective hashrate; further, other people can pull off scaled-up mining operations in the short term".
It sounds very easy to say "just get 51% of the world hashrate" (you seem unaware of what a feat this is) when in fact worldwide many actors would simply mine to the extent they can get away with it, probably including large centralized operations, and prisoner's dilemma and lack of worldwide coordination of bad actors would quite likely prevent such a 51% attack. I also don't think you can make the presumption the decentralized, down-scaled nodes will be a minority of the total hashrate if up-scaled mining becomes difficult (of course, it is possible, no in is denying that a 51% attack has always been a means of attacking bitcoin).
> but the reality is that it's primarily hydro power because that's the cheapest. Thing is, in regions with cheap hydro power, the miners can use so much power they end up making power more expensive for everyone.
Err that depends.
The thing with dams is that the water can never rise above a certain level, else the structural stability of the dam is compromised. So often they have to run the dam with turbines disconnected to lower the water level since there's nobody on the receiving end.
That's also why some blocks of hydro power are very cheap (and time sensitive).
The renewable argument never made any sense to me. Electricity is mostly fungible if crypto wasn't using it something else would or it would be used instead fossil or it wouldn't be produced in first place. Every single one of those options much better than spending it on fairly useless speculative bubble asset...
> escaping capital controls in certain currencies). Some are not (ie all the illegal usages).
I don't really understand how these are different, given that escaping capital controls is by definition illegal (otherwise you could just use the normal banking system).
> I really wonder what happens to Bitcoin (or any PoW coin) when it no longer becomes economical to mine new coins. Does all the computing power just move on to the next coin?
We will find out exactly this relatively soon when Ethereum moves to Eth2 and Proof of Stake.
You don't have to control what software is permitted to run: you can just ban changing USD to Bitcoin and vice versa. Stop the exchanges. Most people aren't mining their own coins anyway.
You can ban a lot of things in life.
Theft. Drugs. Crime in general.
You may not understand how the OTC market works in jurisdictions like China that do indeed ban cryptocurrency exchanges and prevent fiat<->cryptocurrency exchanges. A fist full of RMB can be swapped for USDT, which can be swapped for BTC from OTC brokers.
What you are proposing can only be stopped by breaking computers, breaking arms/legs and breaking economies. Violence is never the solution to problems.
I never suggested that it will completely erase cryptocurrencies. I do think that it will help making it less popular, which would be great. I sure think that while making theft illegal didn't stop it, it reduced the number of theft cases.
> (2) No person shall manufacture, import, offer to the public, provide, or otherwise traffic in any technology, product, service, device, component, or part thereof, that—
> (A) is primarily designed or produced for the purpose of circumventing a technological measure that effectively controls access to a work protected under this title;
> (B) has only limited commercially significant purpose or use other than to circumvent a technological measure that effectively controls access to a work protected under this title; or
> (C) is marketed by that person or another acting in concert with that person with that person’s knowledge for use in circumventing a technological measure that effectively controls access to a work protected under this title.
I think you can make a valid argument against me with this. DCMA is a hell of a law.
To my defense, I would argue that this speaks to production and sale of circumvention tools, and does not actually criminalize the execution of this code. But I weaken my original position about outlawing software. Thank you.
Not all crypto is proof of work based. People have and are inventing better systems that reduce energy costs and provide more utility. “crypto” does not deserve a unilateral ban.
Difficulty of the work needed is lowered if it takes too long for miners to find the solution. I'm sure there is a point at which mining becomes not worth doing financially. But that may not happen anytime soon.
Actually Bitcoin and a lot of other coins have a coin limit (to prevent hyperinflation) and once the limit is set than the miners would make money on the transaction fee.
Let's not commit the fallacy of comparing unrelated big numbers. How does the energy and time per usd transferred compare between btc and the traditional banking system?
I did a quick calculation. If you assume all of the U.S. energy consumption goes into stocks and forex trading (the only numbers that were easy to google quickly), dollar trades are about the same dollars traded per kWh as bitcoin, maybe a tad worse.
In actuality, there are more dollar transactions that doesn't include, so the numerator is bigger. It also uses much much less than the total U.S. energy consumption, so the denominator is much much smaller.
A bigger numerator and much much smaller denominator tells me that bitcoin is much much less efficient than traditional systems in dollars traded per kWh.
Off the top of my head, I could see it as reasonable to pay a 5-10% cost in energy efficiency comparing with traditional transaction systems to enable a truly decentralized model. Bitcoin is nowhere near that in terms of cost/transaction.
I was going to do a very fancy calculation but 136 TWh would be perfectly ok if Bitcoin did at at least 5k transactions per second with a transaction fee of $0.10.
I think parent is advocating doing folding@home over stuff like PoW crypto since the former advances science and the latter (handwaves vaguely) doesn't.
Unless someone manages to turn out a hash collision or an all-zeros hash (spoiler: they won’t), there is absolutely no scientific value in the PoW process itself. (Whether the resulting cryptocurrency has scientific value is left as an exercise to the reader).
These “assets” are usually just participation trophies in decentralized Ponzi schemes. Bitcoin is propped up by the ongoing tether scandal that is now larger than Bernie Madoff ever was.
I'm sorry you don't find it useful. I do find it useful, in fact millions more like me do. Usefulness is a subjective measurement.
If you think we are crazy, that's OK. Bitcoin is opt-in. You can choose to ignore it, you can think its fucking useless, but in my opinion, you are wrong.
How do I opt-out of the CO₂ exhaled by the 7 billion other humans on the planet, or the CO₂ from Facebook or TikTok data centers (as IMHO both are very harmful for the planet)?
The claim being put forth is that crypto's environmental cost FAR outweighs its utility.
Your counterargument is that such concerns are unraisable simply because one can avoid personally using BTC. Or to put another way "if WE destroy the environment it's none of YOUR business"
And that is wrong -- the atmosphere is shared by all. Changes to its composition are everybody's shared concern.
If you want to argue that some other business's activity are too damaging to the environment to be allowed to continue... fine, make those arguments. But don't bat away concerns about crypto's absurd environmental cost as "none of your darn business"
> in regions with cheap hydro power, the miners can use so much power they end up making power more expensive for everyone.
Not if there are laws to keep electricity price within a range. In places like Canada electricity is not necessarily traded in public for-profit markets, and must by law stay below inflation. I agree that in can be a problem in (most of?) the US.
> I really wonder what happens to Bitcoin (or any PoW coin) when it no longer becomes economical to mine new coins.
That might be the equilibrium price, the electricity/production cost. Isn't the pow algorithm smart enough to keep rewarding at least above electricity cost?
If you force the energy price to a certain subsidized level, then bitcoin farms will buy it all, masquerading as homes and small businesses, until you will be forced to shut down some large industrial consumers, putting real people out of work and making everyone poorer.
There is simply no way to waste valuable energy that's somehow imune to social and ecological effects.
And while I wholeheartedly agree with the crypto/GPU issue, all in all, the GPU market is overall fairly tiny compared to TSMC's total output.
Most of the GPU issue is linked to massive demand and conservative allocations that were bought by GPU makers a while back. Adding extra allocation because of demand is really hard specifically on constrained high end nodes like GPU uses.
Car makers and others in the industry don't use those nodes, they use older ones and those are equally constrained.
The GPU/crypto issue is the cherry on top of an already very constrained TSMC, it seems.
> And while I wholeheartedly agree with the crypto/GPU issue, all in all, the GPU market is overall fairly tiny compared to TSMC's total output.
From an economics perspective, it doesn’t really matter. The price of a market-rate good is the marginal price of producing one more unit. As long as cryptocurrency is financially incentivizing people to consume 100% of the marginal production, prices will continue to go up.
This also cascades down their lineup: Vendors who would like to use 7nm for their chip will instead choose 10nm, which will bump some of the customers who wanted to use 10nm back to the 12nm or 16nm nodes and so on.
I think you may be massively underestimating the lead times on those decisions. We're talking years.
Also, you can't simply choose another node last minute, your design is intimately linked to the PDK of the node. Porting to another node is far from trivial.
Yes. We've seen chipmaker MSRPs inflated, especially in the most recent releases like the 6600XT. Chipmakers are constrained by contractual obligations and PR though. AIBs have also inflated their MSRPs for cards. They seem to have contractual agreements with the chipmakers to sell some portion of their cards at Chipmaker MSRP, at least around the launch. However, to their best ability they are putting their die allocation into their highest end card variants that have the highest MSRPs. Then distributors, retailers, and PC builders mark up the cards over the AIB MSRP. Again, these are often limited to some extent by agreements with their suppliers. If all of these markups don't reach the market rate the cards are bought up by scalpers with bots and sold at the true market rate on eBay.
Nvidia already added in a hardware crypto mining block to it's latest models and they are still having massive supply issues. It's not about crypto even though everyone likes to make a big deal out of it to feel better.
The latter does not follow from the former (even if it might be true). The mining block only affects the latest batches of a few models, but profitability of mining also affects demand for the other models, for AMD GPUs, and in the second-hand market. Unless the supply of mining-blocked GPUs is sufficient to cover all displaced demand from these other segments, it doesn't necessarily help much.
They don't block all mining, just specific algorithms. It's easy enough to mine some other coin and convert to the one you want. In fact, that's often more profitable.
A recent Typhoon (one of the half dozen that hammered China) refilled all their reservoirs about a month or so ago.
The one thing that complicates economics is that it's not a linear equation like they teach in Macroeconomics. The equations are more or less "true" in a vague sense but each term is actually a differential equation term in both time and other quantities.
If you are familiar with simple one variable ODEs per STEM undergrad courses, imagine the 2nd order forms but now those exist across many variables and time spans. It makes you realize that CNBC is just pulling causes/effects and making predictions out of their asses.
> A recent Typhoon (one of the half dozen that hammered China) refilled all their reservoirs about a month or so ago.
Yes I should have phrased that better, sorry. The thing is, making one wafer of chips takes roughly 3 months, and everything else has even more terrible lead times. So there's a delay in the impact of events.
> One thing I heard is that the drought (worst one in 50 years in Taiwan) has really affected them ...
This is the one aspect of manufactuing of chips that many are actually ignorant about - it needs huge amounts of pure water.
This is the main reason why India doesn't yet have any fab, despite having the human resource for it and investors and government showing great interest to setup one.
Two or three decades back Intel desired to setup a fab in India, but the deal fell through because it would mean India would have to divert a huge amount of water from agriculture and potable water for the population to such fabs. TSMC has approached the indian government many times (even this year) seeking permission to setup a fab in India. But the indian goverment has remained non-commital.
The indian government believes it is in their long-term interest to wait for fab tech to evolve to use less water. Indian states already squabble over sharing water resources, and India has to feed a billion people, and such a diversion of resource doesn't make sense in the long-term.
(Perhaps the James Bond movie was right - we'll all be fighting for fresh water in the coming decades. :)
This is a very false premise. Water is no reason to not have a fab, just like Japan builds semiconductor & nuclear plants in a seismically active zone. So does Israel in a dessert.
India receives way more rain than it requires even in the drier parts, the problem with soil degradation and deforestation that came with unchecked "modern" agriculture.
The productivity & economic benefits from a FAB and semiconductor ecosystem in India will far outweigh any negative, putting a plant close to a nuclear plant and a massive desalination plant to ensure both power, water as well as security will do the job.
> The productivity & economic benefits from a FAB and semiconductor ecosystem in India will far outweigh any negative, putting a plant close to a nuclear plant and a massive desalination plant to ensure both power, water as well as security will do the job.
Diversion of water resources to fabs (they require around 2 million liters a day; Intel fabs consumed 9 billion gallons / 3.4 trillion liters in 2015) can gravely affect our food security as 60% of our cultivable land are in "drought prone" areas. Moreover, the water released from the fabs is toxic and can heavily pollute our environment even after they are treated and stringent standards are complied with (and expecting that is a joke in India).
Your suggestion of using a nuclear plant and a desalination plant might seem practical, but it ignores the fact that the byproduct of desalination plants are huge quantities of various salts. And dumping them causes pollution (if you dump them back in the sea, it can create dead zones for kms around as most aquatic life cannot survive in such a "salty" concentrated environment).
The run off from agriculture and tanners is several orders of magnitude greater.
If you factor the over all benefit to the pollution and resource consumption, a fab is several orders of magnitude better, why else would everyone else scramble to get one built.
A fab at the expense of food security is not a good deal. When you can't feed your people, you are at the mercy of other nations, and that greatly harms national security. India used to once have frequent famines, and had to beg food from other countries. Today it can feed its billion+ population because it is the world leader in producing fruits, vegetables, grains, diary produce etc. A starving population doesn't care about CPU chips.
What happens to the water the plants use? Is it polluted and unusable, or just tied up temporarily and can be released [safely] back into wherever it was diverted from?
> The manufacturing of semiconductors generates wastewater that contains heavy metals and toxic solvents. Due to the hazardous nature of these liquids, semiconductor companies historically have incurred clean-up costs related to groundwater contamination of up to USD 100 million ... Some companies are proactively meeting these challenges and mitigating related risks by reducing their reliance on chemicals and pro-actively improving the treatment of water before discharge. STMicroelectronics, for example, recycled and reused 45% of the total water it withdrew in 2015 and it partnered with Micron Technology to install on-site fluoride waste treatment facilities.
The report also gives concrete numbers: Intel fabs consumed 9 billion gallons (around 3.4 trillion liters) of water in 2015.
Apparently a fab roughly consumes around 20,000 tons (around 1.98 million liters) of water per day:
> The chipmaking industry is water-intensive. According to nonprofit China Water Risk (CWR), an average semiconductor factory consumes roughly 20,000 tons of water a day, equivalent to the daily water use of a city of 58,000. Yet despite its high water consumption, the chip industry—in Taiwan, at least—is not normally a noticeable drain on communal resources.
Evaporation, goes to ground water table, stored in plant matter, goes back to river now enriched with fertilizers from fields. If there is fight over water the last one is likely quite small fraction.
Is the returned water safe though? That's what I wasn't sure about, do the fab plants pollute the water to the point it's unusable, or is it able to be re-used(drank).
> One thing I heard is that the drought (worst one in 50 years in Taiwan) has really affected them,
The power issues earlier this year in Texas also impacted production.
Weather issues are often the cause of the bust cycles in silicon manufacturing. Just people outside of the industry rarely discuss when memory prices or whatever spike due to a typhoon hitting Japan/Korea/Taiwan. This year, we've have major weather events hit chip production across the globe, at basically the same time, plus the pandemic.
I agree that it is simply an excess of demand, the whole bitcoin rant though. Really?
There is a demand for semiconductors that exceeds supply. You have made a leap that the current demand driver is bitcoin and its consumption of GPUS. As much as I appreciate the sentiment that Bitcoin is a waste, that leap doesn't hold up when you look at the actual numbers for semiconductor demand.
The number one demand item for 7nm semiconductors is cell phones.
If you read the various analysts reports you will find this sums it up well: It(demand) is mainly driven by the increasing demand for consumer electronics and the integration of semiconductor into several devices across sectors like automotive, industrial, communications, military, and others. (that is from this report[1] that for $2500 you too could read it (https://www.marketreportsworld.com/global-semiconductor-mark...))
During the pandemic it appears that everyone upgraded their TV (at least in the US) which, not surprisingly, consumes a lot of semi-conductors. The pandemic has also botched a lot of logistics which will sort itself out but these small bumps can cause unusually large ripples in the supply chain because so many industries are built around "just in time" delivery of raw materials. Semiconductor manufacturing is a process that is intolerant of frequent line starts/stops.
So TSMC, knowing that new fabs cannot come online instantly, has chosen to risk alienating their customers a bit by raising prices to capture the value associated with their assets (productive plants) during this period.
[1] To be clear, I didn't pony up the $2500, I looked at a number of sites that reported on the contents of the reports and pulled quotes from them. Market information is itself a market.
I love how crypto has become easy-to-blame, when in fact datacenter after datacenter is being stuffed to the brim with chips for ML workloads. It's important to recognize when consumers are being squeezed by multiple market forces.
The issue is that crypto mining is the only system designed to have negative marginal utility for each GPU consumed. It literally pays people to take energy and GPUs away from the open market for zero incremental gain to the network.
The difference is that each additional iPhone or ML accelerator or cloud server CPU provides incrementally more utility. More ML accelerators means more ML processing.
Proof-of-work mining is the only system designed to get less efficient with each additional unit of compute power. Add another GPU to the mining network and the network will automatically become less efficient. More GPUs doesn’t mean more transactions, but the system pays people to essentially burn these GPUs.
That’s the complaint. Not that chips are being used, but that the marginal output of each additional GPU is negative.
I would argue that most ML processing by most "big data" companies also has marginal negative output for each additional GPU. Certainly anything to do with social media or adtech and a lot of analytics in general is a net negative for the world, while consuming electricity and processing hardware.
Right, the ML processing is competitive in nature, just like the crypto mining is. Much of it also rests on illegal activity that just hasn't been discovered yet, whether overtly illegal (like botnets sloshing around ad money under fraudulent pretenses) or just mild to moderately illegal, like prima facie unauthorized database usage or use of scraped data. There are also uses of data that are illegal if it comes from US citizens and ?? if they are non-US citizens. Just because the government is collectively a fat slug when it comes to law enforcement in these areas doesn't mean the laws aren't on the books.
You can't seriously be conflating all the money laundering and ransomware attacks enabled by crypto with a tiny bit of crime which happens at the margins of ML (like it might in any industry) can you?
It structurally pays less per GPU as you add more GPUs, because the security budget is a zero-sum budget. (though, that's only if you ignore extrinsic effects, and some people argue that more security = higher coin price = higher security budget).
Even if the security budget is zero-sum, the amount of actual security added to the network with each additional GPU is constant.
It seems like your argument is that cryptocurrency needs to use over half of the computing power on earth at any given time to consider it secure. That's a very bad standard. Visa and MasterCard don't need this much computing power for security.
Security is an easy argument to reach for when you want to justify something stupid.
Even those things you have decided provide marginal value employ people, and you never know where innovation will come from. Crypto is comparable to a ponzi scheme and is an asset with almost no real value, trading only on the greater fool theory.
They are not being reductionist. Each additional hashing unit cancels out at the next difficulty adjustment, having a net effect of zero on the overall working of the network.
Yes, the hash rate of the network increases, thus the difficulty for an attacker to perform a double spend. But, arguably, that's an already insane level of resources, many orders of magnitude higher than what a bitcoin-like payment system would reasonably require.
So even if you consider the network socially useful, pouring thousands of times more hashing power that the network actually requires to perform its task securely is still a waste of resources.
The amount of value extracted by the person mining reduces as more compute power starts mining, but that's actually a negative feedback loop that will reduce the total spending on chips for crypto mining.
The amount of security provided to the network is constant in the number of GPUs. As you add more GPUs, the network gets harder to attack. Does the network need as much security as it's paying for? Nobody is really sure, it's an open debate in the crypto industry.
I really don't see how the current hashrate is required to prevent a 51% attack, that seems way beyond what is necessary.
These Proof of Waste systems aren't really a negative feedback loop, ultimately asset prices determine mining yield. Capacity will keep being added to the network until the difficulty rises such that you are required to destroy $99.99 of resources for $100 of crypto. Until that point it's worth buying more hardware and energy as you get a positive expected return.
>The issue is that crypto mining is the only system designed to have negative marginal utility for each GPU consumed. It literally pays people to take energy and GPUs away from the open market for zero incremental gain to the network.
That's because the utility is not in processing transactions but the 'security' of the blockchain, which is funny considering the most of these blockchains are still looking for a use case that justifies their market cap and yet still run into performance limitations processing transactions.
I have capital I want to deploy. Zero/Negative interest rates are a reality. Traditional broker managed investments are a hot zone for fees. Newer tools for access to markets is covered by reddit day traders; real estate is also highly competitive due to airbnb businesses and other forces.
Capital will always find the easiest path, and right now that's crypto and no one on HN is complaining about micro-VC activity.
It's mostly not ML either. There were production hiccups do to covid, increased consumer demand for chips due to covid, increased demand of mobile chips because the mobile industry is still massively growing, among other issues.
Crypto and ML are each rather small aspects of the issue at hand.
But datacenter CPU/GPUs take the same wafers, machines, and factory time slots as consumer GPUs. Which that is the price increase this article is about.
Ok - I wasn’t replying to the article - I was replying to a comment about crypto-induced price increases of consumer gear being blamed on datacenters, which just isn’t true. It’s not that OEM prices are going up, it’s that retail prices are. Nvidia might sell a consumer GPU which is immediately resold on the retail market at 100% markup.
> Nvidia might sell a consumer GPU which is immediately resold on the retail market at 100% markup.
The same might be said of event ticket resellers making a big markup.
But the reality is that, in the case of event tickets at least, it's very common for the original sellers and the resellers either to be secretly the same party, or at least have some kind of profit sharing agreement.
It's simply that the original seller doesn't want to be seen to be price gouging, so doesn't want to put the list price of an item up, but still wants to gain. That's frequently done by it being very expensive to become an 'approved seller' or similar.
Agreed. I actually don't have a problem with ticket resellers for the very reason that if there is an active resale market the tickets were too cheap to begin with (or more likely, as you note, were never really available at those prices anyway as the promoters are selling at 2x face through the back door).
If someone is willing to pay 2x MSRP for a GPU then the MSRP is wrong (or rather, it's just a "suggestion" as the S indicates). My only point is that up until recently, there was not a long-term secondary market for GPUs. That is a new phenomenon that has been driven by crypto (as opposed to a big reduction in supply or huge increase in PC gamers).
It's still mediated by the same bottlenecks right? NVidia would be able to produce more consumer GPU's if they weren't competing for capacity at TSMC and Samsung
Not just ML, with everything moving to cloud services, while consumers still use phones plus laptops, the number of chips used per person more and more.
I'm not an advocate for cloud services by all means, but aren't they generally much more efficient in using their resources than traditional on premise solutions due to the sharing of the underlying infrastructure?
Are they also raising prices of CPUs? Arm chips? Your argument is based on an assertion that sounds like it could be right, but is it? Hasn't every country in the world been injecting cash stimulus into people's hands in an epic attempt to boost demand across all industries? Also why are miners the target here. Are there not deep learning, gamers, anything else who also want to use these monsterous compute units? (GPUs specifically). Raising the price is temporary because someday another chip maker can be born. Every day the lithography tech is getting cheaper and cheaper is just a matter of time till this resolves itself. And let me ask, are chips a human right like water and food? We're getting close but honestly life can go on without GPUs can't it?
Miners are the target of anger not (necessarily) because it’s wasted energy,[a] but because them using GPUs creates a self serving cycle: Difficulty goes up, more GPUs are bought to counteract that, repeat. This isn’t helped my the monetary reward for doing so.[b]
Deep learning, gaming, etc. are different in that someone doing 2 way SLI doesn’t make others need to do the same. If my online opponents are playing on RTX 3090s, I’m not forced to upgrade to play with them.
[a]: The energy usage is a byproduct of the proof-of-work system
[b]: hence why things like Folding@home, SETI@home, etc. never gained massive popularity like coins did; they don’t reward you except in feel-good thoughts
> using GPUs creates a self serving cycle: Difficulty goes up, more GPUs are bought to counteract that, repeat. This isn’t helped my the monetary reward for doing so.
Not quite. Price continues to go up, courtesy of Tether, which makes the mining rewards more valuable. This encourages new miners to join the network which pushes difficulty higher, until the difficulty level rises such that every unit of output costs more to produce than its worth. Then it's no longer profitable to mine.
But again, Tether has no choice but to keep things pumping.
Honest question: is it possible to make a cryptocurrency that does not rely on some real-world scarce resource (e.g. chips, electricity, drive space)?
This seems to be a cornerstone of what makes any crypto valuable in the first place (scarcity), and as long as that scarcity has real-world consequences, it seems like it would have the same result.
Chia claimed to "Save the world from proof-of-work wastage", but it would just turn a GPU crunch into a HDD crunch.
Decentralized currencies rely on proof of economic waste.
It's possible to imagine forms of economic waste that have different externalities, e.g. different environmental impacts.
But we don't have evidence yet (that I know of?) that introducing one decentralized currency reduces waste in another one -- if you did introduce a new currency that wasted economic value in a new way, there's no evidence that crypto as a whole would waste less of the existing things, it would just start wasting that thing too.
In this model crypto is like ... an emergent antipattern of networked civilization, where the network starts to waste every economic resource that it can "see" and provably waste.
>In this model crypto is like ... an emergent antipattern of networked civilization, where the network starts to waste every economic resource that it can "see" and provably waste.
Exactly. As I've said to a number of people I've discussed crypto with, I'm happy to advocate for making PoW systems illegal. They'll happily gobble up a Dyson sphere of energy and a solar system of hardware in service of running a payment system with a tiny fraction of the throughput of any credit card company or bank.
They are a complete waste of resources, energy, and human endeavor and society would be better off without them. I'll reserve judgement on PoS and other non-PoW systems until they're more widespread and I actually understand them better.
I think that's right. Like I said I reserve judgement until they're more prevalent and I've learned more about them. I have read that Ethereum's PoS implementation will use <1% of it's current PoW energy.
Proof-of-Stake? I thought that PoStake still cranked out blocks at a fixed rate (say, every few minutes), and presumably the block size could be increased to increase throughput, if not latency.
>Chia claimed to "Save the world from proof-of-work wastage", but it would just turn a GPU crunch into a HDD crunch.
Chia actually hasn't resulted in an HDD crunch. The economics of it are such that further investing in HDD storage isn't profitable, yet already Chia is the most decentralized and resilient crypto. I say resilient because price collapsed but allocated storage hasn't.
The underlying problem is that you need to prevent sybil attacks, i.e. group X creates many peers in the network and so gets much more voting power than they "ought to" have.
The only known way to do so is through attaching a real-world cost to ~votes. The way PoW does so is through equating votes and hashrate, so votes cost real money because hashrate requires real hardware. PoS works in mysterious ways, but the basic idea is that votes are equal to prior balance by essentially putting some of that balance "into escrow" in order to vote.
Sure, but a lot of the pro-cryptocurrency people say (or used to, at least) that cryptocurrencies will liberate us from the existing power structures. But it doesn't really, because the people with pre-existing power can just as easily buy their way into power over cryptocurrencies too.
How could a rich existing power exert control over who and how someone uses bitcoin? There is no mechanism by which such control could be exerted.
The worst you can do is monitor traffic and try to correlate wallets and transactions to known identities and products similar to attacks on Tor, then use good old regular meat space violence. Privacy oriented coins like Monitor make this even harder, to literally impossible, too.
I suppose you could start mandating some central registry for wallets or some such and refuse to process any transactions not originating or destined to a registered wallet, but that would require a hard fork and people to willingly go with it.
Technically you don't even need miners, computers, or electricity to exchange cryptocurrency. You can just exchange scraps of paper with a key phrase printed on it that says "wallet contains X btc". The walled could have been emptied before the exchange and that would still require verification by looking at the chain, but otherwise it's perfectly doable, and literal bitcoins do exist.
Well, yes, of course you're right, but there's a world of difference between "technically" and what the reality is for most people.
Between what the EU seem to be doing and the prevalence of centralized exchanges, it seems like what you're describing is happening. But my point was really just that you could still buy yourself into a position of control, either by controlling the hashing power or otherwise. Nothing changes that.
Note I'm not anti-cryptocurrency, I'm just wary that much of the hype is unfounded. Maybe in theory based on correct things, but the reality isn't so clean. Personally, I'm hoping a privacy-oriented DeFi coin will take off. There are a few contenders in this space to watch.
You're still not printing over 21 million bitcoin though. And you're still not deanonymizing transactions (for properly designed privacy coins). I think monetary policy and privacy were the two main arguments for crypto in the early days before we started funding every startup with a half-baked "XXX on The Blockchain" idea.
The real question is how do you create a pool of voters who is 'representative' of the users.
One way is you could look to some historic list of people (like for example, everyone who had a US passport issued in 1992), and say they are the voters.
An evil actor subverting more than half of those people would be very hard.
That's essentially what the Ethereum PoS transition is supposed to look like, isn't it? Take a snapshot of the PoW chain, boom, here are all your initial tokens/coinz eligible for staking to get votes.
“Work” doesn’t have to be wasteful. It could be based on the number of pieces of litter that were collected, the number of meals to the homeless that were served. It’s just easier to measure when work is something wasteful.
If BTC mining relied on pieces of litter being collected you'd get people spending $millions on factories to produce litter and vehicles to distribute it just so more could be collected. That sounds crazy to type out, but it's no stranger than the situation we currently find ourselves in. The most advanced tools man has ever created are being used to create chips with no practical value, but that are good for converting energy into tokens through pointless calculations.
I don't believe it is possible to create a proof of work system that is not a "proof of waste" once the reward is high.
Spare CPU cycles felt like a waste, that becomes dedicated processing, that becomes dedicated hardware, that eventually became designing custom hardware.
Kim Stanley Robinson proposes this in his recent novel, The Ministry for the Future. He mostly skips the details around how it would actually work though, beyond needing an enormous amount of human labor for verification. In real life, that would break the decentralization of it.
While Crypto is certainly a contributor to demand, I think a large part of the problem is related to a demand spike similar to the 'toilet paper shortage', where companies try to stockpile because they hear there will be a shortage, which in turn makes the problem worse. Additionally, COVID stoppages caused chaos in supply.
The toilet paper shortage is kind of a bad example, because it was caused by a very real and sudden shift in demand from the very seperate commercial and personal toilet paper manufacturing markets.
So is GPUs for crypto mining demand, though. GPU makers don't hate it because they have a conscience, they hate it because it can collapse any moment when a bubble unrelated to their product bursts (or someone makes an ASIC, for some of them).
BTC GPU mining hasn't been profitable for years. ETH GPU mining (the main one) has gotten way less profitable recently and is scheduled to not be a thing in 6 months to a year. The GPU demand for mining has already lessened in the last months and you can see decreased 2nd hand prices..
Every comment that blames BTC for chip shortages is likely largely incorrect and looking in the wrong place.
TSMC has Bitmain as a low-priority client, and the first one to be cut when demand is high like earlier this year[0]. Bitmain itself is having even bigger reduction in orders to TSMC currently[1] so again, that's the wrong place to look for increased chip demand.
At any rate, I was mostly responding to claims about 'GPUs are profitable for mining' in the parent.
This could be a possible explanation for the Fermi Paradox. Advanced civilizations almost become space faring, but at around the same time they discover cryptocurrencies and funnel all their planets resources into it, until they become stranded and die out.
There is an ugly feedback loop scenario where the crypto bubble accelerates inflation due to the high demand crypto places on energy and semiconductors and the proportion of global GDP downstream from energy and semi. Inflation then loops in more demand for crypto assets (as an inflation hedge) creating ugly feedback loop until governments act to break it.
I highly doubt the chip shortage a la increased chip demand from a company perspective has to do with gaming/mining GPU demand.
Much of the supply demand can be traced to increased car chip demand (Toyota cutting production by 40%) as well as general device chip demand across many sectors.
Cryptocurrency miners are handing TSMCs boatloads of money that they are using to grow their R&D efforts. Cryptocurrency, in the long term, will be amazing for the advancement of GPU technology.
(though we should still move crypto away from GPU mining for environmental and security reasons)
The state of cryptocurrencies is nothing more than a visible symptom of larger systemic issues that have clearly gotten out of control. It would be much more beneficial to try and understand why the market has decided to consider it valuable in spite of everything you have said rather than just blame and dismiss it based on its impact. We aren't paying the price for cryptocurrency, we are paying the price for policy that has made cryptocurrency attractive.
GPUs are a marginal factor in this. The main factors are several impacts from Covid-19, US trade restrictions on SMIC chip exports from China (a big supplier of low and mid-range chips which this is actually all about), and the Taiwan drought.
The Covid-19 impact itself is multi-pronged. Many chip plants closed or reduced output due to the lockdowns, but also the silicon used in chip production is the same material used to manufacture vials so vaccine production also impacts high quality silicon availability. Finally early in the epidemic a lot of manufacturers that consume chips, such as car makers, cut production. This lead to some chip foundries being closed as they were non-viable and it's taking a long time for those foundries to get re-commissioned or other foundries to expand to make up the gap in capacity.
Nevertheless any impact on production would set them behind their production targets which would need to be caught up. Even if it didn't impact production, it would certainly have impacted costs, hence contributing to the price hike.
> This is yet another way we’re all forced to pay the price for cryptocurrency.
Bitcoin has existed for more than a decade. This is yet another way we're all forced to pay the price for COVID-19.
When everybody started working from home, companies whose employees had desktops had to buy them all laptops. People with old home computers bought newer ones once they started using them 14 hours a day instead of two. Demand went up.
A bunch of car manufacturers thought the pandemic was going to crush demand for cars, so they canceled all their chip contracts. Then it didn't and they bid up the price of chips trying to source all the ones they canceled the contracts on.
The primary cost of Bitcoin mining is the electricity, not the hardware. If there is more demand for hardware, the world can build more fabs. Assuming we're given some advance notice, which this year we haven't had because COVID-19.
And the reason the electricity consumption is bad is that we don't have a carbon tax, so people are burning coal to mine Bitcoin, which is Very Bad. But it's also Very Bad that people are burning coal to power electric heaters or elevators or anything else, because the problem is burning coal.
If people were mining Bitcoin by building wind turbines, nobody would care. It might even improve the world by increasing the economies of scale for renewable energy production.
But I haven't seen "buy bitcoin here" signs in the cities(that aren't tech hubs) until recently. Now people are randomly asking "what's the next big crypto?" trying to make quick money.
People like come up with "Bitcoin uses more electricity than <small country>!" as a shock headline, but that's just statistical chicanery. A very small percentage multiplied by the whole world is a large absolute number. It's like saying "the global paper industry uses more electricity than all of Australia!" People think of Australia as being big, but as a percentage of the whole world, not that big.
The primary advantage is that cryptocurrency mining is highly price sensitive, so if you have something like wind turbines that produce more at night even though there is more demand during the day, it can be profitable to only mine cryptocurrency at night. Which increases the revenue for the wind farm and makes it more profitable, which means we can have more wind farms and fewer coal power plants. But only by like a half a percent of the power grid, because that's how much is going into Bitcoin.
Is it tragedy of the commons, then? Once things settle, won’t people learn that adding 1 additional GPU only increases my cost by 1 and doesn't increase my reward? Can’t we design a system where the number of transactions processed scales with the compute power of the system rather than striving to always keeping block time constant?
Not talking about other more volatile crypto as you'd be tempted to assert their values should be 0 in a perfect world.
You can argue all day that's wasting energy, the crypto-ecosystem, chip manufacturers, miners, energy providers, users, developers, traders, bankers and companies driving billion of dollars of revenue (Aka value for human beings who will produce and consume goods in the real World with this revenue ) out of it are proving that you're missing the bigger picture.
I've never understood HN's concerns with crypto.
Following the same logic, people are literally burning petrol in their cars while it could be used for <insert supposedly better use case>
Is it for ecological reasons ?
You're still moving 2 tons of metal everywhere you go, and eating meat.
Cryptos are not less moral or valuable than plenty of things you do or cherish. Let people mine crypto, eat meat at Wendy's driving their F150, because the world is not black and white.
"Shortage" is a technical term to describe a situation where outside influence is preventing supply from meeting demand at the market clearing price. I do think crypto falls into this category, as well as numerous government policies (every tax creates a shortage).
well, frustrating maybe. But you know what, chips production is not like gravity, there are human controlling who gets what chip... Someone, somewhere could better arbitrate things...
And it's not the invisible hand which is, as often, well, invisible.
Don't forget the main advantage of mining over buying coin is to avoid all fiat associated regulations. So it's not just about slightly cheaper tokens, it's really about distributing the cost of tax evasion and money laundering.
What makes you say this? Nigeria, a country with a currency that loses purchasing parity power every year, has the highest per capita usage of Bitcoin in the world.
Sounds like the people that need it ARE using it...
I'm saying that Nigeria has per capita the largest Bitcoin using population on Earth compared to other countries, which is a counterpoint to the argument that people that need it (escaping local currency controls + inflation) are not using it. I am arguing that people that need it ARE using it.
If one of the biggest upsides to crypto is that it can be used as an end-run around authoritarian regimes, then a policy proposal that doesn't hurt this while curtailing some of the bigger downsides gets some points in its favor.
I am hoping that someone comes up with a semi-reusable proof-of-work system , like chess board positions, Mersenne primes, or protein folding. Kind of like Re-captcha.
GPUs are not used to mine Bitcoin, they are used to mine Ethereum. Bitcoin is mined by different chips, also made by TSMC, but in far smaller quantities.
Cars cost tens of thousands of dollars, so I think they should be somewhat entitled to use whatever tech they want. I’m more upset at every fridge and light bulb and oven having WiFi.
It's no fun before supply catches up with demand, but when it does, we'll be in a world that has higher chip production capacity and maybe some advances will come from all that extra money chip makers will be making. It seems more promising long term than, some under-produced stagnant technology like nuclear power plants.
1. This is specifically mainstream node, or everything except leading edge 3nm / 5nm.
2. Those price has been like that for sometime since the chip shortage. They just make it official now.
3. The mainstream and social media will and always have been blaming it on TSMC.
4. Everyone wants all the capacity they need, while not paying a cent more. This announcement will "somehow" made it fair for everyone in the industry.
5. The current component price has very little to do with wafer pricing. As with all commodity pricing, they will only swing back once supply catches up.
6. In general I think this is a good, TSMC has been leaving far too much money on the table and benefiting GF and SMIC.
> 1. This is specifically mainstream node, or everything except leading edge 3nm / 5nm.
Yes, but only because leading edge is contracted out well into the future. Prices for those nodes will definitely be higher when contracts are written.
Gosh I wish Apple with their loads of cash and drive to have the highest performing chip would get big ole fabs on US and EU shores. The fact that a disputed country owns the best fabs is downright scary for the future of the world (add in that the disputing country is the next superpower).
TSMC is building a $12b manufacturing facility in Arizona that will be able to produce the 5nm chips used in Apple products. They are fully aware of the geopolitical risks and are taking responsible steps to mitigate them.
To the extent that the higher unit prices help them cover those investments, I'm more than happy to pay whatever portion of the increase gets passed on to consumers.
There's a complex game being played here by TSMC, the Taiwanese government, and the US. TSMC as a company probably wants to be resilient by having manufacturing in multiple locales. But it has a unique set of benefits provided by Taiwan - a solidly educated workforce with fewer options than in, say, the US for top-tier jobs, a dense and well-connected country that allows them to shift talent much easier than elsewhere, and a government thats extremely interested in maintaining their dominance. These factors make it hard to move out of Taiwan. Taiwan itself wants to keep its golden goose, the one thing it has that guarantees it a seat at the negotiating table with the US, China, and EU. Why voluntarily give up the one thing that makes people want to defend you, especially in light of NATO promises of protection falling flat recently with Afghanistan and the anti-internationalist angle made prevalent by Trump. The US, of course, wants to protect its economy, reliant on high-tech components and be the guarantor of a country that gives the US leverage over other countries.
The 5nm plant is a good step for the US, but AFAIK, produces a small fraction of total 5nm output. And by the time it opens, 3nm will be the leading edge. It feels like a small concession by TSMC/Taiwan to keep the US off their back, but not significant enough. I mean, how do you even transfer thousands of talented engineers that far, with logistics and factory construction, too? It's not easy.
Apple know that just setting up fabs without all the infrastructure to package, test, and then manufacture products with the chips doesn't really solve anything.
The last time Apple tried to do any manufacturing in the US the project failed because they couldn't find anyone to manufacture the screws: https://www.nytimes.com/2019/01/28/technology/iphones-apple-...
"I was hoping things would settle but it doesn't appear to be the case."
I've come to the conclusion that it's not going to 'settle' for a long time. Not until competition arrives from new foundries (several years out, although they _are_ being planned and built) and competition from Intel and their GPU platform appears and eventually becomes competitive. So your wait will be a couple years longer.
Proof Of Stake with Ethereum won't change anything. There are many other POW block chains around and people invested in GPUs farming will migrate, find profit and continue indulging their livelihoods. COVID isn't going away and isn't really causing any of this to begin with; it's just the go-to excuse for every complaint.
The real problem is obvious to me; the manufacturers, particularly those in Western countries, have been on a 20 year long lemming-like off-shoring race to Taiwan and S. Korea foundries and now there is a foundry oligopoly and competition no longer exists. Prices go up. Big surprise. And it won't return to 'normal' anytime soon.
With that in mind I bit the bullet ordered up all the parts for my new machine two weeks ago; $3k with a 3070 TI GPU. About a week after my order the the exact same GPU appeared in stock from the exact same reseller with a $200 price hike. The prices aren't going down. They're going up.
Update: that GPU is about $250 more now than what I paid on 8/6.
The worst part of all of this is now my PC is this precious asset. Used to be a GPU could die and "meh, whatever... could have used an update anyhow." Now I feel the need to pay for a 'protection' plan.
The good part of just paying the cost is now I'm no longer involved in 'newegg shuffle' nonsense or participating in 24/7 Live Stock Watcher videos on YouTube[1]. How messed up is that?
Everybody saw Intel slacking off a decade ago including Apple and Google but Google didn't have the guts to start their own CPU design Department despite being the #5 largest computer maker on Earth!
Very true. But they have hedged that very well with long-term contracts, so they are effectively first in line for most of their needs. I would guess that this is still biting them on some components (they have so many in so many products), but they appear to be riding through this pretty well.
A modest new desktop will be definitively much better than anything you could have had 11 years ago.
Unless you are looking for high-end gaming, the most important parts will be your motherboard and SSD (nvme) speed. You could get all high-end components except for the graphic card, and just wait it out and replace it later when the prices come back.
Try finding a PC starter bundle that includes a gpu from GameStop or some other retailer and you can find one for MSRP. The trade off is you’re buying other PC parts you may not need, for example a PSU.
Onboard graphic is pretty good these days, unless you plan to game high-end stuff. I'm actually using my 4790k's onboard graphic (2014) as a daily driver since I use my GPUs for compute only.
Prices for used GPUs are also inflated, though I think they have started to come down a bit. Once Ethereum moves to proof-of-stake (early 2022?) the market will probably get flooded with used GPUs.
I’m saving up for the flood of old GPUs. I can finally realize my dream of taking all my ML work off the cloud once I scoop up some old mining GPUs for relatively cheap
You really may be waiting a good while. And if a desirable used model is ever a decent price, it’ll be a unit burnt out from mining fake money that’s only used to buy real money or a mail order of LSD.
Mining usually does not "burn out" a GPU. Most miners cool their cards adequately and undervolt them. If a GPU is somehow damaged from mining then it’s typically observable - it either does not work or it artifacts quickly, making it easy to detect and avoid.
Even if they do that, they’re still running it 24/7/365 for possibly years at a time. Wouldn’t that be some serious wear and tear? Genuine question - I have no clue.
Most of the wear and tear on a GPU is going to come from temperature changes due to the workload changing, or, for the fans, the card spinning the fans at start up, ramping up and ramping down due to a change in load, or halting the fans at shutdown.
It's very possible that an undervolted, well-ventilated card, run with a steady workload 24/7/365 for three years would have _less_ wear and tear than a card used for gaming and then shut down every night for three years.
I don't think proof-of-stake will actually work. Not because of any technical reasons, but because 95% of the people involved in crypto just see it as a free money printing machine. Once their hardware is no longer useful for mining, they will just move to some other proof-of-work crypto.
It's incredibly unfortunate that Bitcoin and similar proof-of-work currencies blew up and remained hot commodities. Because I do think that cryptocurrency is abstractly a very exciting idea, but the popular implementations are fundamentally abysmal.
If proof-of-stake works out, particularly as a currency people can actually use, then maybe nobody wants to buy Bitcoin anymore and the whole thing dries up as only speculators are left.
Just to be clear, Bitcoin isn't creating any GPU demand now; mining ASICs are too good. Ethereum is also becoming ASIC-dominated, but being scrypt-based reduces the ASIC advantage.
If many miners move to another cryptocurrency, the difficulty for that network will go up, to the point that it's no longer profitable to mine it (electricity costs exceed income). Unless the price for the cryptocurrency goes up as well, but it won't magically do that just because of an influx of miners. After all, those other networks don't have the utility that Ethereum has.
This. I have a gut feeling that a lot of mining rigs were bought with credit and haven’t paid for themselves yet. Those rigs will find a new asset to mine rather than being dumped. Outlawing mining is the only thing I see that causes prices to move downward
Like how China banned it in May, and it has been creeping back?
You can't ban mining. At least not in any jurisdiction with free speech type laws.
Be careful what you are asking for -- the ability to run arbitrary software should not be controlled by governments.
My current GPU is a used 980, it's mostly ok but it makes noises when doing certain tasks for some reason (like scrolling a browser window), I'm really not sure what causes it but it's put me off buying second hand cards, if it fails there's not much you can do to fix it
I use 980Ti too - i'm lighting 6 candles on the cake this year...It keeps me warm at night and during the winter.
I am surprised it's still working after 6 years of daily use with that TDP and heat released.
My 970 is still limping along... kinda hilarious that it's paired with a 5950x in my new workstation. I'm impressed it can still play modern games pretty decently on my 1440p monitor although usually I've gotta dial it down to medium gfx settings.
Gonna try to hold out till next gen, crossing my fingers supply issues have been sorted out by then.
I’m running an RX 470, since I upgraded one of my monitors to QHD, it has been having some issues. Luckily, I hate action games, don’t care much about high-end graphics, and grew up with 30 FPS as target ;)
Some day I’ll be able to upgrade to a new budget GPU.
Forgive the naive question but why doesn't the US government fund several foundries in the US immediately? $20B over 3 years is a drop in the bucket compared to a lot of military spending and these chip shortages surely have downstream impact on the defense industry, no?
At this point I'd be much happier subsidizing chip manufacturing than most of the other ways this country spends money overseas.
Might scare Taiwan and embolden China? A huge reason we hold Taiwan as an ally is their chip production capabilities. If we become self reliant on those then we don't really need Taiwan and China might try to invade them, which would be bad for our own national security interests.
Well, for one, we don't want a humanitarian crisis and civil war or even international war... that's bad for the whole region and thus bad for US interests.
We also don't want China able to just take over land and territory at will, which they've threatened to do all over the region. That destabilizes the region and once again is bad for US interests. This also emboldens other nations with expansion desires which threatens stability across the world... No US ally would trust us again if we just let Taiwan be taken over... again, bad for US interests...
There are 190 other countries that applies to so where are they?
I don't share the ego for the US' hegemony or I also don't place importance on the US' soft power. I do care about semiconductors though.
Is there a stronger argument, because I'm still stuck on what specific strategic advantage that Taiwan has to the US, after we duplicate their one interesting company's infrastructure.
Its obvious that isn’t important to me, right? This sentiment crosses party lines.
We uniquely have the resources to be the most developed nation, domestically, but subsidize other countries’ defense instead and they subsidize other advances in their own country that we domestically deem too expensive.
So how about we get this semiconductor geopolitical nightmare out of the way and use our massive domestic resources the way we can. There are consequences, not the US’ problem.
There are a lot of people thinking like me, so hawkish people will need more articulate arguments for specific US foreign policy with specific countries.
> We uniquely have the resources to be the most developed nation, domestically, but subsidize other countries’ defense instead
...and in return, the rest of the world plays ball when we want to advance our interests. The basis of foreign policy is I'll scratch your back if you'll scratch mine. Our protection of Taiwan shows other countries that we will put ourselves on the line for our allies. It practically forces us to go to war if China invades because we'll lose access to our all-important semiconductors.
Key word forces. If we lose our dependency on the semiconductors we're no longer forced to go to war. Which means China can try to invade and might succeed if we deem it unnecessary to protect Taiwan because "fuck you, we got ours." By keeping ourselves in this vulnerable position there's an implicit trust and understanding that we will be there for them.
> There are consequences, not the US’ problem.
Funnily enough this is the attitude of many citizens under protection of US hegemony. I think you will receive more varied responses from the Ukrainians and Taiwanese.
It is one thing to build massive fabs it is quite another to run them profitably especially with China in the game, you are competing with a large pool of cheap R&D engineers and a captive market.
This was a mistake with 5G, where each country guarded their patent pool, R&D and failed to collaborate with strategic partners.
Lots of governments have spent billions of dollars trying to build the next Silicon Valley. Shenzhen and arguably Taipei succeeded, but all the others failed. It's not as simple as just spending money; the people whose knowhow makes the endeavor possible have to decide to go there, and then they have to be in charge of it.
The Manhattan Project would have failed, as the German project in fact did, if Oppenheimer hadn't been a physicist, and if the US weren't full of Jewish refugees from Germany who just happened to be physicists. Its immense budget was necessary to its success, but not sufficient.
$20B is enough for 1 foundry if you can find the people but the people at Intel do not make good hires because they don't know how to make state of the art chips at Intel anymore! Our only other domestic fab are the former IBM fabs owned by global foundries in East Fishkill New York but global foundries has given up on state of the art chips too!
Supply shortages are the excuse du jour, but so long as TSMC has the kingmaking node I expect it to become expensive. Very expensive. Expensive enough to move margin from the king their node made right back to TSMC.
Right. If AMD and Intel both got out of the fab business, on account of it being low margin, then TSMC would promptly yoink their collective margin, because if you choose the king you get to charge for the privilege. Fortunately Intel decided to compete instead.
See also: AMD and NVidia, and whichever other pairs rely on leading-edge nodes.
Well, the higher the price, the more incentive other companies have to catch up and compete with them. (Or take a more flexible portion of the market by price.)
Chip fabrication is a difficult market, with barriers to entry only similar to stuff dominated by governments, but there is still a little bit of competition there.
That depends on if someone else gets in. AMD or Intel could if they wanted to - without the expertise it will be hard, but a few expensive years of research and they can get back into it. Hire a few TMSC people.
Or it could be someone completely new. The problem is hard, but solvable.
TMSC does have incentive to keep prices low enough that that doesn't happen. They also have other incentives to treat good customers well (though who the good customers are is an open question)
I am not an expert, but I thought it was more like a few expensive decades than years. Intel is already in this business and they have only been able to fall farther behind TSM even given billions in investment.
>a few expensive years of research and they can get back into it.
I haven't kept up much in the last couple of years, but back when I did, I was under the impression that AMD couldn't actually afford to do this. I know their condition has improved with the good fortunes of Zen, but is it enough that they could get back into semiconductor fabrication?
Sure, but it's not not a price hike on the leading node. Technically you are correct, the leading node price hikes will happen separately, behind closed doors, but it makes sense to talk about them here. There won't be a HN post about the non-public price hikes because, well, they aren't public.
Bitfinex (which has been outlawed in the USA) keeps printing Tether to prop up the price of Bitcoin, graphics cards are snatched up as soon as they are manufactured, and they're using an Argentina-sized amount of power to keep the whole thing going. Yes, it works, but no one on earth will ever use these Bitcoins to buy two pizzas ever again. What's the end-game, at this point? As many bagholders as possible?
Not for Bitcoin. PoS is a broken system that reinvents oligarchies. Proof of Work is the only secure protocol that works today -- possessing the real, desirable properties like decentralization and censorship resistance.
Don't tell me about vaporware ETH2.
The end goal for Bitcoin is Dyson spheres around stars. And it can be no other way.
You might be joking, but I am not.
When put on a Kardashev scale, all these energy concerns seem pathetic. Terrawatts? Lol. Zettajoules of energy await us in the stars!
I agree that terawatts seem pathetic when you're thinking about even Kardashev Type I, but I would hate to see those zettajoules of energy spent on zero-sum competitions instead of goods and services that improve our standard of living. That's still true even if the competition consists of Bitcoin instead of the nuclear arms race, Wheel of Fortune, Democrats vs. Republicans, World War II, and NASCAR. I mean, at least Bitcoin and NASCAR don't burn cities, but they aren't building space habs either.
0.1% of our energy being spent on Bitcoin mining is fine, 1% might be a little worrisome, 10% would be a real wakeup call, 50% would be a crisis, and 90% would be ruinous. Is there a natural limit? Competition from alternative currency systems with cheaper transaction costs, maybe?
If you don't want to hear about ETH2 (which is, for the record, not vaporware), Cardano has been PoS from the start and is the #3 crypto [1]. I wouldn't be surprised if it overtook Eth soon.
Yes, I think the informed Bitcoin investor is well aware of the value that decentralized peer-to-peer censorship resistant electronic money brings to the table.
I care. Everyone I know cares. The core developers care. The ecosystem cares. I think you are radically underestimating the value of such things.
I'm pretty sure most of them just see mining as quick easy cash. If people actually valued the decentralization aspect then more would be using it for day to day transactions rather than just speculative investments.
Miners? Not a chance. They know exactly what they are doing for decentralization and censorship resistance. It requires large capital investment over many years, and often includes long energy contracts with providers. If you believe mining Bitcoin is 'quick easy cash' I am afraid you are mistaken (in 2021).
Some Retail investors? Sure -- but I'm sure you could find misinformed investors in all fields.
The point nocoiners miss is that holding Bitcoin IS using it. It is storing value over larger periods, and especially through jurisdictions with poor fiscal policies. Many like me will not sell their Bitcoin at any price -- especially not for dirty fiat.
This might be the start of the end of high-end PC gaming, unless prices get back to a more reasonable level. Consoles (and phones) are very competitive and cost significantly less.
Or perhaps it just causes the games to have lower performance requirements. Is integrated graphics the future of PC gaming?
High-end PC gamers were always about money-is-no-object gaming. None of the tech they use has ever been strictly necessary.
The vast majority of gamers still use 1080p. More people login to Steam on a Mac than login to Steam with a 4K monitor attached.
Maybe the high-end is more expensive now, but you still see forums riddled with people paying $1000+ for cards. If nobody was paying for it the price would go back down. High-end PC gamers are the most price inelastic gamers on the market!
I'd say that integrated graphics are the present of PC gaming. eSports games, Minecraft, Fortnite and similar games dominate sales and play time, and they don't have any serious requirement for discrete graphics.
This is all to say that the high-end isn't going to just go away. The overall gaming market is bigger than ever, so high-end gaming along with other niches like VR are here to stay. These niches on their own are just large enough to support a variety of gaming markets, while in the past they might have been too obscure to warrant investment and development effort.
Depends on what you mean by high-end. If you want to run the latest AAA games then you will need a decent machine. If you want to do >1080p gaming then the requirements are even greater. But it's actually not that expensive, if you play 2 every day for 8/10 days, then you will play 584 hours/year. If your gaming setup costs $3000 every 5 years, that comes out to $1/hour. That's a lot cheaper than most hobbies.
The popularity of eSports is because pretty much anyone can play those games not that high end gaming is becoming a niche. Plus eSports & low-end gaming are super popular in less wealthy, non-western countries where as high-end gaming is going to be limited to developed countries. Within developed countries, high-end gaming is not a niche like VR.
I think it's interesting that we have three options now for sidestepping this by streaming games and yet none of them have made much of a dent.
Microsoft seems to have the best business model but you're stuck streaming the games on Game Pass, and their latency seems to be the worst in my experience.
Stadia seems to have the best tech in terms of latency, but the weird "it'll only work on these certain Chromecasts" launch situation put me off it for a bit, and then there's the business model - trusting Google to honor purchases on this service in perpetuity.
Luna is the one I haven't tried but seems to be going for an approach where they partner with publishers to sell catalog packs, even if they're not quite there yet.
The availability of the high bandwidth, low latency FTTP internet is lagging severly enough for these to have no impact.
The use cases where I, a high end gaming PC owner, would use streaming services are: travelling. Except unless you visit a city center, you aren't getting 5G, and you aren't getting WiFi 6 speeds. I'm in the UK, and here my parents live, where I live and where my friends live, none have the speeds capable for it. I imagine things are even worse in 95% of the US too.
Game streaming won't happen in a country unless most of the country has 100MBit down
I feel like high-end PC gaming, trying to maximize graphics of brand new titles has always been a silly dick-measuring competition.
For me the sweet spot of games is 2-8 years old. Reviews are out, so you can skip the crap. Patches have been released to fix bugs. And everything that has stood the test of time runs pretty well on low- to mid-tier hardware by now.
Dust off those games in your steam library that you never touched before.
Stadia doesn’t work “fairly well”. It’s also low-res (still) compared to a local GPU, which may be okay for some but certainly isn’t sufficient for serious gamers.
Yes it does? What issues do you have with it? It just works.
> It’s also low-res (still) compared to a local GPU
1080p for free or (upscaled) 4K for 10€/month is fairly decent, and on the latest games is better than most integrated or older GPUs.
> serious gamers
What percentage of the gamers' market is "serious gamers"? For the people that "must have" 8K low latency etc sure, cloud gaming isn't the best option. But IMHO they're the minority and vastly more people are "casual gamers" that are more than fine with playing 1080p/upscaled 4K for peanuts for multiple years ( how much does a RTX 3080 cost, and how many years of Stadia can we buy with the same amount of money?)
I had a gutfeel around 3080 launch it was going to be big, so grabbed one on launch day (had to wait for months for the local store to have stock though..). Now seeing them 2x the price. Crazy times.
I am in need of an upgrade but having a harder and harder time justifying the expense for what is a hobby.
I guess eventually my system will be old enough that I will be better off just ordering a prebuilt, where I'll get gouged a bit overall but not by some insane amount just for a GPU
That was my assumption - I managed to pick up a 3080 TI FE at MSRP in the UK a couple of weeks back. I would have gone for the 3080, better price/perf, but since the only ones available are AIB for the same price, or more, than the 3080 TI FE that made the decision for me.
The chip shortage is interesting to me. In that it seems to be mostly driven by bad forecasting both by clients and the foundries. But the foundries seem to get nothing but upside out of it. Perhaps it swings next time and there's excess capacity.
There are some reports of collapsing lumber prices, when everyone was talking about the sky falling and prices skyrocketing two months ago. https://financialpost.com/commodities/lumber-prices-collapse... So your idea that semiconductor prices might fall a lot in the future seems reasonable.
I've heard the industry is full of exaggerated orders now to, everyone ordering more than they need, to see what they can get. Not a great way to have efficient trade.
It will swing. It always swings. This isn't the first time this has happened.
People see these nice, tasty prices, and bring more capacity online. That takes time measured in quarters to years, but when it comes online, there are suddenly a lot more chips. Prices fall, but nobody will reduce output, because they have to pay for all this new capacity they just built.
It will come. It's going to be a rough year or two, though.
Food availability is driven by weather as much as anything. The North American wheat harvest has been poor this year so the price of wheat is rising around the world.
(tweet text, also includes a video and some more info)
"With @Bitnob_official's Lightning Network integration, we now have free, instant, non-reversible remittances of any size to and from the United States, Nigeria, and Ghana.
$10 in my US bank account became spendable NGN for @bernard_parah in seconds.
The real value is only in the USD and NGN. Using Bitcoin to move the value is only propped by hype-fueled insanity that doesn't need to exist. International banks could easily create an international money-transfer system that would be far more reliable and secure than any decentralized Ponzi scheme. An in fact, such systems do exist - for example I used to earn in USD and spend instantly in other countries using international Visa/Mastercards like Revolut or Wise.
You are lucky that you had access to those services. In many poor countries the majority of citizens do not have a bank account.
Even if international banks created these services (why haven't they done more of this already if it's easy?) these people would not be able to access it.
I don't know a single person who uses bitcoins or a single store that accepts them... and it's very risky for companies due to it's extreme volatility (the price may increase/decrease by hundreds percent within days or weeks)
sarcasm? Nobody uses crypto for remittance payments to poorer countries. It is mostly used by rich folks in countries like China to get around exchange restrictions.
> The company has said it will make $100 billion in capital investments over three years through 2023.
>At the same time, "it will be very costly to manufacture in the U.S. and Japan," a member of the TSMC executive team said. The company is building a cutting-edge semiconductor facility in the U.S. state of Arizona and is considering opening its first Japanese chip plant in Kumamoto.
TSMC heed the call of US to build a new fab in AZ. It seem logical to raise their prices to meet this "demand". A fab in the US will not generate as much profit as Taiwan or China fab due to labor cost/benefits.
Chip making uses "a lot of water." (Parts of) Arizona is experiencing water shortage, particularly aquifer draw down.
I wonder what the effect of chip making in Arizona will be in Arizona, relative to other water use.
Parts of Arizona by law don't meter aquifer use, which disincentivizes current ag users to conserve, and has incentivized corporate ag to relocate there. (From memory, and I may have botched that memory)
I wonder if TSMC is locating in an unmetered location, or getting a formal water regulation break.
"Counterintuitively, the famously thirsty industry can even improve the local water supply due to a focus on reclamation and purification—Intel has funded 15 water restoration projects in the Grand Canyon State with a goal of restoring 937 million gallons per year, and it expects to reach net positive water use once the projects are completed."
Theoretically that will spawn some real competition which is sorely needed. In reality fabs are hard to spin up, but this does draw attention to the matter.
Contracts are what they are, but they don't preclude discussions like "if we honor this as-is, it leads to a situation where we have no incentive to value you as a customer".
Firm orders years in advance are in general far more valuable than much more expensive orders today. With the years in advance orders you can make plans around them.
That might be the reason for rumours of higher storage (1tb). An excuse to charge 30% more to not only offset increased costs but to increase profit margins.
Intel has their own foundries, but they are using TSMC N6 for their announced gaming GPUs, so they are subject to much of the same supply constraints. Apple is also using TSMC for their A-something chips.
Intel was stupid. They could have become a foundry 20Y ago like every other chip making company on Earth but they chose instead to greedily try and make cell phone chips three times and they failed all three times at the cost of many many billion dollars! And now nobody really wants their chips and AMD chips are selling like hot cakes using up all The foundry resources while intel chip manufacturing lines stand idle, and useless because Intel thought they would be on top forever!
At the moment, we'd pay several X over price versus not being able to find needed ICs at all and having to redesign our satellite for ones that are available.
The government of Taiwan has been willing to maintain the status kayfabe, including maintain the "Republic of China" name and holding China's seat at the UN for a couple of decades. It all seems quite ridiculous from the outside, but a lot of people in the region still get quite pissy about the particulars.
The official name for the country is 中華民國, or in English, "Republic of China". Its national anthem is called 中華民國國歌. "Taiwan" 臺灣/台灣 is not found in the official name of the country. Sometimes it uses the name "Chinese Taipei" in international organizations.
I’ve heard that action alone triggers an advance from the CCP
More likely it triggers more lipservice but people havent been willing to risk it (nor has there been broad enough consensus in Taiwan to relinquish their own claims to the name and the mainland and the same other territorial expansionist policies that the CCP has, people seem to forget that)
Inflation is a rate of change, not an absolute change. When people say inflation is transitory, they mean that the rate of change is transitory (prices will stop going up). But the inflation that has occurred will not go away. You will not pay less when the inflation subsides. You will merely stop paying more.
This is what happens when you print $5T and refuse to let the economy go through any sort of healthy contraction.
How big chunk chips are of the median consumer's basket? Sure, everyone and their dog has smartphones, but not everyone has to buy the next-gen high-end ones. (Where the price increase happens.)
The inflation has been there for a while now; wages have been artificially depressed for decades while the stock market was growing 15% a year. But when there’s inflation for the rich we call it an “asset bubble”.
Is TSMC based out of Taiwan? Do they have fabrication plants anywhere else? What company will pick up the demand if/when China takes control of Taiwan?
Samsung and Intel are the other two closest fabs in terms of technical capability. TSMC is clearly in the lead right now. And while the gap is very large in technology cycle terms (they're between 1 and 2 generations ahead at this point), in real world terms (like... considering what a fucking gongshow China invading Taiwan would be), it's not -that- big of a deal. It'd be a major speedbump, but not world changing (as opposed to the other ramifications of an invasion).
TSMC has a few smaller fabs currently outside of Taiwan, 2 in China, 1 in the US (WaferTech not the Arizona plant under-construction) and 1 in Singapore.
Designers are getting overwhelmed because they're having to redesign boards for components that are available, only to realize by the time they finish the changes, the part they redesigned for is also out of stock.
There's been some talk of a Chip-induced cold war, and I would say that's not crazy given the market conditions. I also love the guy who posts here all the time saying he works for the company that makes the machines that makes the chips and they can't make more machines because they don't have the chips that go in them. XD