This Dumb Industry: The Disaster of GPU Prices

By Shamus Posted Tuesday Mar 13, 2018

Filed under: Column 127 comments

Last week I gave a lightweight overview of what bitcoin is and how it works. ) Like I said, people can make money by certifying blocks for the blockchain. And to do this they need to try lots and lots of SHA256 hashes, looking for the magic nonce that will complete a block. The more computing power you have, the more hashes you can try per second. This improves your odds of winning the SHA256 lottery and netting yourself a few thousand dollars worth of bitcoins.

This works to keep bitcoin secure, but it has also created an intense and sudden demand for lots of computing power. At some point these bitcoin miners looked at consumer-grade graphics hardware and realized that they were an incredible deal in terms of cost vs. computing power. So they began buying them up and stuffing them into minimalist cases where the cards can sit and crunch numbers all day. They’re not even hooked up to monitors! They’re just crunching on SHA256 hashes!

(Note that when I say “bitcoin” I’m sort of doing this handwave-y inclusion of crypto currency in general. I’m talking about bitcoin because it’s the most recognizable and notorious of the currencies, but a lot of the things I talk about below apply more to the other currencies than to bitcoin specifically. But I don’t want to clutter this article up with all the asides and asterisks it would take to sort out the various strains of cryptobux. Just remember that this article is more concerned with graphics hardware and less concerned with crypto.)

The Graphics Hardware Market

OUT OF STOCK. Which is fine, since I can't afford any of them anyway.
OUT OF STOCK. Which is fine, since I can't afford any of them anyway.

All of this has pushed prices up. Normally a new graphics generation comes out, and then the price drifts slowly downward as the technology ages. Eventually a new card comes along to replace it, and the old one undergoes a price drop. Most of us budget-conscious gamers shop in this “recently dethroned” spot on the price curve, which is where you get the most bang for your buck. Any newer than that and you’re paying the hefty enthusiast markup. Any lower than that and the power falls off more quickly than the price.

Unfortunately bitcoin miners noticed this sweet spot too. So they bought up all the best cards. And when those ran out they bought all the next-best cards. (In terms of money vs. power.) And then the next-next best cards. The supply dried up just as demand spiked. And still the miners want more. They sell graphics cards in 6-packs now!

How bad is it? The GeForce GTX 1070 launched with a suggested price of $380. A year and a half later and you might expect the price to have dropped a little, but instead it’s at an insane $1,000!

And that’s assuming you can find them in stock.

Hoping For The Crash

If you're going to mine for bitcoin, this is the pickaxe you'll be using. Bring several.
If you're going to mine for bitcoin, this is the pickaxe you'll be using. Bring several.

The one thing PC Gamers can hope for is that the bottom will fall out of the bitcoin mining market. I realize it’s not nice to hope for the misfortune of another, but shortages have a way of making us turn on each other.

Everyone kind of has the sense that cryptocurrencies are going to take a tumble at some point, but nobody can agree on when it’ll happen or how far it’ll fall.

Note that I’m not saying this means cryptocurrency is “Doomed”, or a dumb idea, or that it’ll vanish entirely. There are reasons people created this stuff in the first place, and those reasons will continue to exist regardless of how the currencies perform on the open market.

I don’t understand all the reasons people are drawn to cryptobux, but here’s one example: Someone looks at the debt their country is racking up, and they decide (correctly or not) that this is unsustainable. Historically, when a country is on the brink of insolvency (or even over the brink) they begin printing money to satisfy their creditors. This fixes the short-term problem and allows the government to keep doing government stuff, at the expense of destroying the value of their citizen’s money. Someone worried about this kind of outcome might start looking for a way to buy something with their money before it becomes worthless. Historically, gold or other precious metals have been popular for this. But gold is prone to theft and it’s difficult to buy common goods with it. But cryptobux could let you transfer your money into something that won’t crash when your government turns on the printing press, and you can trade it over the internet. You can’t buy a lot of things with bitcoin yet, but all it would take is for someone like Amazon to look at all those billions tied up in bitcoin and decide they’d like to capture some of that value. If they (or another retailer of similar stature) began accepting bitcoin as payment, the entire landscape could change. Bitcoin might go up even more, which would further fuel the mining market and make everything that much worse. (For us PC enthusiasts, anyway.)

There are other reasons to be interested as well. Tax evasion, money laundering, privacy concerns, lack of trust in traditional global banks, and a host of other worries drove demand for this alternative money. It doesn’t matter if you think these people are wrong, stupid, crazy, or immoral. They believe this money is needed, and I don’t think those beliefs will change anytime soon.

However, it’s also true that a lot of people riding the cryptocurrency bandwagon aren’t here for the privacy, freedom, security, or whatever else people think bitcoin can give them. They’re here because they want to get rich. Their demand has inflated the price, which has drawn in more people, which has further inflated the price, and so on.

Here is a bitcoin mining computer. It's an open-air case with SIX(!!!) graphics cards.
Here is a bitcoin mining computer. It's an open-air case with SIX(!!!) graphics cards.

Which means that we’re in a bubble not unlike the Dot-Com bubble. In 1998, investors thought that the internet was the future of commerce. And they were basically right! But the rush of money attracted speculators, copycats, and get-rich-quick types who dumped their money into unwise ventures and produced a bubble. Certainly that’s at least somewhat true of bitcoin today. There are people buying Bitcoin solely for the reason that it’s been going up. They don’t care about alternative currencies, they’re just jumping on the bandwagon. Once it starts going down they’ll bail out, and the bubble will pop.

That might be a relief to us PC gamers who are terrified our old, obsolete cards are going to fail because we can’t afford to replace them in this market, but the truth is that I don’t think the markets will correct overnight. Our woes are not tied to the value of bitcoin, but the value of bitcoin mining. Even if bitcoin crashes tomorrow and is suddenly worth one-tenth what it is today, that’s not going to magically solve our problems. Maybe miners will stop building new machines and supply will gradually recover. But how low will the price have to go before the miners turn off the server farm and put their graphics cards on the used market?

(Then again, mining apparently puts an incredible stress on hardware, since it runs the GPU at full load 24/7. Most people don’t game 24/7. And even when they do, very few games really max out the hardware. So a card bought a year ago for crypto mining could easily have the equivalent of a decade and a half of wear and tear. Which means it might be near failure. If the used market were flooded with a bunch of unreliable, burned-out cards, then people may shy away from used and prefer new. Great. Another thing to speculate about.)

So bitcoin might continue to go up, or it might tank. And even if it does tank, it might not help the graphics card market. And even if it does, it will probably take some time. And even then, the market could be complicated by fraudsters and a glut of unreliable worn-out cards. In the long term this problem will sort itself out, but in the short term I don’t see any hope for people who need an upgrade.

So why doesn’t NVIDIA just raise production to meet demand? Isn’t that what companies are supposed to do?

Well…

Why Not Make More?

Ironically, everyone stayed at work late tonight to mine bitcoin.
Ironically, everyone stayed at work late tonight to mine bitcoin.

Like Apple, Qualcomm, AMD, and several other tech giants, NVIDIA is a “fabless semiconductor company”. They don’t have a facility for fabricating their own chips. Instead, they pay Taiwan Semiconductor Manufacturing Company (TSMC) to make the chips for them. These fabless companies schedule time on the TSMC fab facility, essentially renting a huge manufacturing complex. They compete with the likes of Apple for fab time, so there’s an upper limit on how many chips they can have made in a given year.

On top of this is some additional drama / complexity. Back in 2015, TSMC announced they were going to build a new “mega-fab” installation. That new facility isn’t online yetI’m basing this on the assumption that if it was up and running, there would have been an announcement.. This isn’t surprising. It takes years to construct, staff, and supply a fab facility, and I imagine it gets even harder if you’re building a “mega-fab”. Note that a fab facility isn’t just a magic box that makes microchips pop out. Each facility is geared towards a certain type of production. As technology improves, we figure out how to fabricate chips with smaller circuits. That’s what’s been driving this whole crazy Moore’s Law business for the last several decades. Back in 2007, circuits were around 100 nanometersHow small is a nanometer? One nanometer is to one meter as one meter is to THE SUN. (Give or take a little.). Every few years we refine the process and the circuits get a little smaller. 40nm. Then 28nm. 20nm. 14nm. Last year, TSMC began selling their first 10nm chips.

Yeah, I used this image last week. But I foolishly spent like two hours making this silly thing and now I'm trying to justify burning those hours on a two-second sight gag.
Yeah, I used this image last week. But I foolishly spent like two hours making this silly thing and now I'm trying to justify burning those hours on a two-second sight gag.

Every time the circuits get smaller, the power requirements change. You need to re-think how all that technology fits together if you want to get the most out of it. Which means you need to spend some time coming up with new circuit board layouts. (Or whatever it is the eggheads do.) The point is that there’s an overhead to moving to the next level of miniaturization. There’s R&D time, prototyping, and testing. You engineers need to spend months solving the new heat, power, and spatial problems. You don’t want to go through that every year.

If you’re a company like NVIDIA and you’re trying to maximize efficiency then you want to make the jump to the new fab process, wait for the technology to stabilize, spin up the manufacturing, then sell lots and lots. If you’re always making the jump to the latest technology then your hardware will always be in the rough “early shakeout” period. On the other hand, if you wait too long then your rivals will swoop in and began selling the New Hotness while you’re still stuck on the old technology. So there’s a bit of jousting and brinkmanship going on between companies as they try to minimize the cost of upgrades without leaving themselves too vulnerable.

And of course there’s the ongoing strangeness that NVIDIA and AMD don’t deal directly with consumers but instead the final products are built and sold by other companies.

What I’m getting at is that the whole system is really complicated and we can’t get more graphics cards by just cranking up the speed dial on the graphics card factory. Is there fab capacity available at TSMC? Is that capacity in the right size range for what we’re trying to make? Are the third-party guys ready to manufacture and sell the final product? This is a deeply complex system, and we can’t just plop down new factories like we’re playing SimCity and wait for the problem to resolve itself.

A Thought Experiment

I tried to find stock photos of graphics cards, but all I could find are images of memory and CPUs. The shortage is even impacting stock photography!
I tried to find stock photos of graphics cards, but all I could find are images of memory and CPUs. The shortage is even impacting stock photography!

For the sake of simplicity, let’s brush all of that aside. Let’s imagine an alternate world where NVIDIA makes their own chips in their own factories and builds their own cards and sells those cards directly to consumers. What does NVIDIA do about this graphics card shortage? Obviously they want to make as many as possible, as soon as possible. But how?

Assuming they’re already running the NVIDIA factories at capacity, they’re faced with a dangerous choice. They can spend billions to expand their facilities and hire and train a bunch more workers. It can take months or even years to get something like that up to speed. But then if the bitcoin bubble pops, all those bitcoin-mining graphics cards will end up dumped on the second-hand market. NVIDIA will have spent a fortune to expand just as the market is collapsing and prices are falling. They’ll have tons of new expenses and costs from the expanded production, but demand for their goods will be at an all-time low as the PC gamers feast on the now dirt-cheap used card market. A turn like that can put you out of business. Even if you survive, it can hobble you for years to come.

On the other hand, what if NVIDIA doesn’t expand and the mining market continues to grow? The graphics card shortage is really hurting PC gaming in the short term. We just had the launch of a half-step console generation. If the shortage continues, some of those PC gamers could migrate over to consoles. If we get a few more big-name titles like Wolfenstein II that launch with high system requirements and lots of driver problems, then then PC gaming public may be pulled towards getting consoles because it costs too much to keep their gaming PCs up-to-date.

That doesn’t just hurt your bottom line this year, that hurts it for several years. Fewer people on PC means fewer games get ported to PC. PC ports became an afterthought for a niche market, which means more sloppy ports and even more pressure to get a console. Sure, some of us will stay on the PC regardless, but a lot of people are more fickle / flexible in their gaming habits.

So now NVIDIA needs to figure out how to respond to this spike in demand, and to know how to respond they need to be able to predict what will happen with cryptobux bubble, which sort of makes them another speculator.

And of course, all of this becomes that much harder to think about if you consider that problem in the real world where it takes a long chain of companies to make the final product. NVIDIA probably really, really wants to make lots of new cards as fast as possible, but TSMC? They might not care at all. The GPU market is just part of their business and this bitcoin stuff isn’t really a serious concern for them. Somebody will need chips made in the future, and to them one company is as good as another.

So what is NVIDIA doing? Are they extending themselves to try to increase supply, or are they holding their breath and praying this whole things blows over as soon as possible?

My guesses aren’t any better than a coin flip, but I really hope this gets sorted out soon. I don’t care about internet funny-money, but I’m really due for a graphics upgrade and I can’t afford anything with prices where they are now.

 

Footnotes:

[1] I’m basing this on the assumption that if it was up and running, there would have been an announcement.

[2] How small is a nanometer? One nanometer is to one meter as one meter is to THE SUN. (Give or take a little.)



From The Archives:
 

127 thoughts on “This Dumb Industry: The Disaster of GPU Prices

  1. Nick Powell says:

    To top it off, bitcoin mining is starting to have a noticeable effect on global energy usage. I’ve read that there are other currencies which somehow get around these problems by not requiring so much computational power to maintain a blockchain but I don’t pretend to understand any of this very well so I don’t know how likely they are to catch on.

    In the meantime, I propose game devs work around the issue by changing every graphics option to say ‘ultra’ no matter what it’s actually set to. Then at least I’ll feel better about my aging GPU

    1. Droid says:

      I’ve read that there are other currencies which somehow get around these problems by not requiring so much computational power to maintain a blockchain

      Yes, that would be the proof of work –> proof of stake transition Florian mentioned two comments down. Of course “proof of work” (declaring the blocks that have used up the most computational power to be the legitimate ones) on a large scale is going to require lots of work (duh).

      It’s almost as if making it hard for people to steal your digital product is going to have negative side effects for everyone, including the honest customers *looks judgingly at DRM*.

    2. Matt van Riel says:

      If we’d implement liquid-fuel nuclear we’d have all the energy in the world at no cost to the environment (actually it’d be net positive because we could scrape Co2 from the atmosphere and use it to create synfuels and kill dependence on foreign oils), so miners could use as much as they like with no real problems.

      Frankly, it’d be another renaissance period of massive change and improvement similar in magnitude to the first discovery of electric, or the industrial revolution. It’s also a perfectly safe technology, unlike conventional nuclear we use now.

      But heh, try getting something like that off the ground. First you have to get past the ‘n’ word, especially thanks to rampantly unsafe nuclear leading to accidents like Chernobyl and Fukushima. Then you’d have to get around America’s idiotic ban on reprocessing (specifically for that country, others wouldn’t have the same issue) because these reactors continually re-burn their materials (meaning tiny amounts of waste, like 95% lower than regular nuclear). Then you’d have to get the dumbass energy industry itself to go along with it and convince them that this would lead to insane profits so they’d actually get on-board.

      *sigh*

      Can’t imagine I’ll see that in my lifetime. And probably not in anyone’s, because we’re too busy building pointless crap like solar instead and by the time we realise our mistake… well, the oil will likely run out and we won’t have the means to transport materials around to build new reactors… at which point we have Mad Max and/or Fallout, but without any of the fun stuff of those franchises. Kind of disgusting that that’s the sort of world we’re going to leave behind as our legacy through utter stupidity and lack of foresight.

      Kind of a tangent there, oops, lol.

      1. Droid says:

        People said the same about fusion, and it never took off because, surprise, there were still gigantic obstacles to overcome before you could actually build a functioning reactor with enough efficiency and throughput to be worth building.

        1. Olivier FAURE says:

          You’re saying that like people have abandoned the idea of fusion / tried it and it wasn’t viable. There are still a bunch of very high scale fusion projects that are making progress every year.

          (but yeah, it’s kind of ridiculously hard)

        2. Matt Downie says:

          30 years ago, fusion power was just 30 years away. Now it’s 15 years away. So I estimate it will actually be here by 2050 or so.

          1. Retsam says:

            According to XKCD, that puts it somewhere between “We haven’t finished inventing it, but when we do it’ll be awesome” and “It hasn’t been conclusively proven impossible”.

          2. Cilvre says:

            MIT recently started pushing Fusion more and received a large grant to continue it, https://www.theguardian.com/environment/2018/mar/09/nuclear-fusion-on-brink-of-being-realised-say-mit-scientists

            1. Daemian Lucifer says:

              Even if we achieved practical commercial fusion today,it would still take decades for fusion power to become dominant and replace all of the old coal plants.

          3. Humanoid says:

            SimCity 2000 was right all along!

      2. Viktor says:

        Nuclear just postpones the problem. Any sort of fuel will run out eventually, and with power needs increasing, we’ll burn through fuel faster now than ever before. True renewables(solar/wind/hydro) won’t run out. Yes, there’s more overhead, but I would like to fix the energy problem. Nuclear just kicks the can down the road.

        1. Droid says:

          *non-serious argument about solar, wind and hydro all getting their energy from the sun, which will burn out eventually*

          1. Decius says:

            *serious argument about defeating entropy in the long term*

            1. Droid says:

              *serious argument about that being as clearly science fiction as FTL drives and time machines*

              1. Kestrellius says:

                Well, I mean…we have a few hundred billion years to think about it. If it’s possible at all, I imagine we can figure out a way to do it.

            2. 4th Dimension says:

              Then two options present themselves:
              A) Shonen route: Give a boy a transforming mecha and have him yell at the problem
              B) Avoid entropy by harvesting the angst inherent in the modern Magical Girl shows.

              1. Daemian Lucifer says:

                Third option:We digitize all ourselves,then fly off in the space between galaxies and wait for the new batch of species to populate this one in order to drink all of their milkshakes.

                1. Asdasd says:

                  Fourth option: we invent sentient AI. We won’t personally benefit as a species, sure, but when the AI goes rogue, enslaves us, uses us as pawns in an intergalactic war and then winks out of space-time during the Big Crunch so it can slide back in for the next cycle, we’ll be able to claim triumph over entropy by proxy.

        2. Henson says:

          As I understand it, renewable energy sources don’t produce nearly enough power to come close to meeting our energy needs. You’re going to have to shore up the difference with something else.

          1. Steve C says:

            Renewable energy could supply everything. All renewable comes from the Sun directly and/or indirectly. Hydro, wind, biochem etc all comes from the Sun. We use only a tiny fraction of the maximum of what we could use. For example the USA could supply 39% of all its power just from solar panels mounted on roofs. That’s for the most energy consuming country needs. Other countries need less. Then there’s water, wind, geothermal and proper solar farms on top of that.

            There are both valid and spurious reasons why that hasn’t been done though. I don’t want to get into that. To say it’s not possible isn’t true. Humanity just hasn’t done that because of reasons. We could though.

            1. Henson says:

              I think you made an assumption about what I wrote; I never said that it wasn’t physically possible to produce the energy. There are several things we have the capacity to do but don’t, because of, broadly, human nature. And human nature is something that we are not going to be able to change anytime soon, if at all. Taking that into account, renewable energy sources don’t come close to fulfilling our energy needs, and probably won’t anytime soon.

              Given the above, the use of nuclear as a postponement of full renewable use in the future, as Viktor mentions, is not a trivial step. We certainly could use the time figure out how to make renewables easier to implement.

              I had not seen that study before, though. Thanks for the link.

              1. Alex says:

                As I understand it, renewable energy sources don’t produce nearly enough power to come close to meeting our energy needs.

                I never said that it wasn’t physically possible to produce the energy.

                I don’t think he made any assumptions there…

                1. Galacticplumber says:

                  Oh yes he did. There’s a difference between won’t, or in other words possible but deemed unlikely and can’t.

                  1. BenD says:

                    The human nature point is the main part of the ‘won’t’ argument; you just added it later. That said, reading into Steve’s text it’s clear that he recognizes human nature as the problem. If that was solved, we could use renewables in quantity enough. (But it’s arguably a harder problem to solve — another one somewhere on the scale between ‘impossible as far as we know’ and ‘we’re working on it.’)

                    1. Daemian Lucifer says:

                      But its not just human nature.Even if we decided today that we want to convert 100% to solar,hydro and wind,and even if everyone decided to work for free in building ALL the facilities necessary to make the change,it would still take at least a decade in order to use all of the viable locations,connect them to the places where power is needed,and fully staff everything with trained work force.The places in the world where such alternatives were easy to exploit,like iceland,have already done so.

                    2. Echo Tango says:

                      I think you guys missed my link (below) to an article explaining why we can’t use renewable energy to power our modern world. The TLDR – storage.[1]

                      There’s literally not enough lead (the cheapest) in the world to store all our electricity for cloudy / windless days. You could probably get by with lithium, but that’s orders of magnitude more expensive. Also note that this doesn’t solve transportation. Electric cars are good enough for intra-city driving, OK for short highway trips since you need to recharge in the middle of the day, and do worse in the winter by using energy to heat the battery and to heat your vehicle. Heavy vehicles (semis, tractors, construction equipment) absolutely cannot be run off of even lithium batteries, and would need a wholly different infrastructure to produce liquid fuel.

                      So in general, we could transition to renewable energy if the world wanted to spend a lot of money[1] making smaller changes to our world / lifestyles. Or we could spend far less if we just scaled back our world to use less internal combustion engines – bicycles for everyone, and no more shipping from Amazon Ebay in less than a month.

                      [1] Also, all of the good hydro sites are already in use, so we’re basically talking wind and solar here. Current crops-to-fuel methods are inefficient, costly, and/or compete with food crops.

                      [2] This is on the order of magnitude of the entire output of the world, plus or minus some fraction. You can’t make medium-to-large changes to all of our different infrastructures around the world for cheap.

                    3. Daemian Lucifer says:

                      You dont use batteries to store that much energy. So far the only way we achieved for storing massive amounts of power is in artificial lakes. But I heard proposals for heat and momentum storage, though those are far too inefficient.

                    4. Daemian Lucifer says:

                      Serves me right for posting on the phone.I meant to say CHEMICAL batteries arent good for mass storage.

                      Also,momentum storage technically already exist,since turbines in power plants keep spinning for a while after you stop providing fuel.So small gaps in fuel consumption are not a problem.But a gap of 12 hours,that would be a problem.

                    5. Echo Tango says:

                      The same blogger has another article where he goes over each energy storage option. Batteries are the cheapest storage we’ve got, for amount of energy. Gravity (pumped lakes), flywheels, and other methods are all very weak, so you need to spend a lot of money, volume, and mass, to get any sizable amount, needed to power a city / power-grid through a cloudy day/days. That entire blog has covered pretty much every angle of this, and has calculations for everything, using publically available data. It’s all hard/expensive, and we’ve got a lot of work to do, to transition to any large-scale wind/solar energy source.

                    6. Daemian Lucifer says:

                      That does not cover one of the most important things:how long does the storage last?Artificial lakes can store power for decades,with practically 0 loss*.Chemical batteries rarely survive more than a few years,and thats if you use them in optimal conditions,and not have them be constantly charginh for 10 months,then drain them over 2,then have them accumulate for 6 months,be drained for 4,etc,etc,etc.Not to mention that the water is replaced technically on its own,and your maintenance of the system is focused only on the “casing”(the dam and maybe the shores),while with chemical batteries youd have to replace the whole thing yourself.

                      *There is some loss due to evaporation,changes in water levels over the years,etc.But seeing how big the amount of water is,and how the efficiency of such storage is not that great per liter,those losses are negligible.

                    7. Richard says:

                      “How long can you hold it” doesn’t matter much.
                      If you can store the energy for >7 days then job done, as the worst doldrums on record are about that long.

                      The problem is “How much can you hold?”, and “How does the capacity drop over time?”

                      Batteries can’t hold very much, and wear out very quickly.
                      Li-ion manages around 500 cycles, lead-acid around 350 before the capacity has dropped “too far”.
                      – So your massive system is great initially, but each cell only lasts about two or three years before you have to pull it apart and re-build it.

                      Artificial lakes silt up over time and need regular dredging – and the silt put somewhere of course.

                      So lakes last much longer than batteries, but the energy density is much lower.

                      I read somewhere that the UK could manage to go fully renewable, if they flood Scotland.

            2. Boobah says:

              There are both valid and spurious reasons why that hasn’t been done though. I don’t want to get into that. To say it’s not possible isn’t true. Humanity just hasn’t done that because of reasons. We could though.

              That’s right! I want my flying car!

              Wait, what? That wasn’t the topic?

              You can’t just handwave the ‘valid reasons’ away and then claim that something is possible. Until those valid reasons are addressed and overcome the thing isn’t possible. That’s what makes them valid reasons it hasn’t been done.

          2. Echo Tango says:

            The produce more than enough power[1]; Solar itself is actually cheaper per watt than nuclear, in some states in the USA already. It’s the storage that’s expensive / hard.

            [1] Google around for 10 minutes, or do the math yourself. You could get enough solar wattage to power the USA in something smaller than Nevada. I thought that Do The Math had an article that computes the size of this solar array, but I could only find his article on battery storage instead.

            1. Droid says:

              I took that as a challenge and Wolfram Alpha actually accepted my query without that much fuss.

              The number means “this is how much the total electrical energy consumption is as a fraction of how much energy from the Sun hits the surface of the Earth.

              world’s energy consumption = 0.0000007551 * Solar energy hitting the Earth’s surface.

              There’s some trickery with our atmosphere and stuff, and of course you can’t use all of it or we would all freeze, but there’s a LOT of leading zeros there.

              Formula I used: Sun’s luminosity (all the energy the Sun is emitting) times the relative size of the Earth “in the Sun’s sky” (how much of the light actually makes it directly to the Earth).
              So: Sun’s luminosity * ( (1 Earth diameter)^2 * pi) ) / ( (1 * distance from Sun to Earth)^2 * pi ).

              1. Boobah says:

                Well, for starters the area of a circle is pi * r^2, not pi * d^2 so that’s a factor of four. And the surface area of a sphere is 4* pi * r^2, not pi * r^2, so that’s another factor of four. And both in the same direction.

                That changes the answer from 7.551 x 10^-7 to 120.8 x 10^-7.

                Of course, Earth’s albedo is 0.367, so ~37% of the light is reflected away. That takes me to 191 x 10^-7.

                1. Droid says:

                  You are completely right, I don’t know what I was thinking.

                  On the one hand, this is obviously just a (probably bad) approximation, since there’s all sorts of weird effects going on in both the Sun’s and the Earth’s atmosphere, the Sun and Earth aren’t completely round and the Sun’s light emission isn’t symmetric, so there’s some variance to all this.

                  On the other hand, I’m somewhat confident that the world’s combined power usage is still 4 orders of magnitude away from the power provided by the Sun. In other words, not even (or barely) 0.01% of it.

                  So, that’s some comfort, at least.

                  1. Daemian Lucifer says:

                    Though to be fair a bunch of that power is unusable.The huge part of it that goes into the weather and water cycle can be used with windmills,rain collectors and hydro plants.But the heat itself and a huge part of the specter are unusable(with our current tech).And of the usable part,we can extract at most ~50%,and thats with huge industrial heat collectors and photovoltaics.Commercially,a house covered with photovoltaics or with a single heat collector could extract maaaybe 20% of the energy,on a good day.

                    And of course,seeing how majority of our planets surface is water,you cant really build anything massive there(again,yet),so no big industrial power generators.And lets also not forget that one important fraction of land needs to be covered with plants(which also consume sunlight) in order to clean co2 from the air,otherwise we will be cooked.

                    So while the energy our planet receives from the sun outclasses our needs by several orders of magnitude,we can only use a tiny tiny fraction of it.One day,it could be enough to satisfy all our needs,but that day is not today.

            2. WarlockOfOz says:

              Storage isn’t the only solution. I expect local nuclear and solar with long distance distribution, whether by cable or otherwise, to be a big part of if. It’ll always be sunny somewhere. Yes, this will need investment.

        3. Decius says:

          With tech advancement within the expected curve we can expect the portion of the Earth’s mass that can be converted to negentropy to last until the sun engulfs the planet.

        4. Kestrellius says:

          As discussed, there are no “true renewables”. As I understand it, though, fusion is supposed to run on hydrogen, which is about as close as it gets.

      3. Paul Spooner says:

        Man, I really want to talk about this, but it seems like the whole topic is inextricably embroiled in politics.

      4. DeadleDark says:

        Six hours video about Thorium but it’s worth it
        https://www.youtube.com/watch?v=2oK6Rs6yFsM

      5. Daemian Lucifer says:

        If we’d implement liquid-fuel nuclear we’d have all the energy in the world at no cost to the environment

        Thats an idealistic scenario,but reality is a tad bit different.First,the fuel is more expensive to produce than the solid nuclear one.Second,the byproducts are more radioactive than the ones in solid nuclear fission,so theres more danger to the environment,not less.And third,which is the key part,infrastructure for these plants would need to be built up from scratch,which is EXTREMELY expensive.

        The real reason nuclear power is not the most used source of power in the world 70 years after it started being used is because every nuclear plant had to be built from zero,while shutting down an old coal power plant is not wise when they can still be used for decades.Also,you cant just shut down a huge facility like that,because you need to think of the workers there,the space it occupies,and all the machinery.Getting rid of those usually costs more than just updating the facility itself to be a bit more safe.

        1. Bitcoin stratigo says:

          The reason nuclear hasn’t taken off are more political than practical, involving intense lobbying from multiple sides.

          I’ll avoid the details because it very much is politics, but expense is only a small part.

          The window is rapidly closing in any case, as renewables get better, the need for nuclear is less.

          Coal’s dead though. We don’t need coal for power generation any more. It’s no longer economic and the market is strangling it

          1. Daemian Lucifer says:

            Coal is far from dead.Its still provides about 40% of worlds electricity needs with natural gas being second with about 20%.

            As for nuclear,even at the height of nuclear craze in usa and usssr,when there was huge supply of nuclear both arms ready and fuel ready uranium,only a few nuclear power plants were erected.Its not politics,its simple maths:When you already have facilities that work,cover most of the needs,and are cheaper to maintain and update than to retire,there really is no point in investing more into replacing them.Only when something else shifts the scales,like the current worldwide push towards climate protection,or more importantly*,the rapid decline of fossil fuels reserves,does the need for replacement become pressing.

            *More important for the maths,nothing else

      6. Alex says:

        If we’d implement liquid-fuel nuclear we’d have all the energy in the world at no cost to the environment (actually it’d be net positive because we could scrape Co2 from the atmosphere and use it to create synfuels and kill dependence on foreign oils), so miners could use as much as they like with no real problems.

        If we had liquid-fuel nuclear, proof-of-work cryptocurrencies would just increase the difficulty of finding the nonce (by demanding a higher number of leading zeroes in the resulting hash digest) to ensure mining doesn’t become so cheap that it’s worth forging blocks.

        Bitcoin is horribly inefficient by design, and will deliberately become more inefficient to compensate for any advances elsewhere.

    3. Paul Spooner says:

      So, according to that article, each bitcoin transaction takes about $30 in power to validate. That seems like a really prohibitively large barrier to transactions.

      1. Steve C says:

        That’s not the cost of transactions. That’s the cost of printing money. Compare the $30 cost against printing a sheet of 500 twenties for a better comparison.

        1. Paul Spooner says:

          Well, that draws into question the veracity of the whole article, and thereby all the comments on it. How much power does it take to carry out a bitcoin transaction?

          1. Decius says:

            That’s the cost of all of the transactions in the block, which has an upper size limit based on the memory requirements to hash the block.

          2. Steve C says:

            Actually after reading the article that article is based on, I’m going to walk my comment back. I don’t know what that $30 figure is analogous to.

            The “cost per transaction” seems to be largely made up. Too many assumptions are baked into the numbers I’ve seen. They are concentrated on current and expected costs rather than historical data.

          3. Steve C says:

            I was curious so I did some math of the cost per transaction based on some numbers I found from the source article and a couple of other locations.

            There are ~200,000 bitcoin transactions per day. “40 megawatts per hour and represented 4% of the global Bitcoin network capacity”

            40MWh for 1 day @ $0.05 is $48,000. If that is 4%, then extrapolated out it is $1.2M for the total industry energy costs per day. $438M per year.
            $30 x 200,000 transactions= $6M per day = $2.19 billion per year.

            Those should be in the same ballpark. I don’t trust numbers unless they can come reasonably close to each other. That doesn’t include capital costs. Which would be another fudge factor that might bring it up to $30. I don’t really know. /shrug. Point is, I feel the numbers are a bit nebulous.

            1. Paul Spooner says:

              So, using your 200k transactions/day, and $1.2M in energy per day, that’s $6 per transaction. Still pretty steep, and it will only get more expensive as the nonce length increases.

              1. Richard says:

                For comparison, credit card companies can charge $0.35 per transaction and make a profit.

                That obviously isn’t the whole energy cost of banking, but there’s a whole order of magnitude to play with.

    4. 4th Dimension says:

      The recent “Serbia and their breakaway province Kosovo are bickering over who will pick up the bill for Kosovo’s electric power shortage and have caused the frequency of the entire European grid to drop” was also apparently caused, in part, by the Bit coin miners who have moved to Kosovo since the power is cheap there.

      1. Droid says:

        See, that’s how you write a realistic villain story!

        EDIT: Yes, there is little love lost between me and crypto miners.

    5. Son of Valhalla says:

      I have no idea why crypto-currency is driving gpu prices up so exponentially (maybe should read the article? O.o), but otherwise, when I started choosing parts to build my PC last year, I was apalled that the prices of GPUs that were ordinarily $150 suddenly jumped to $250.

      Kinda sucks, but there isn’t a whole lot to do besides, as Shamus says, waiting for the crash to hit, similar to what happens to any currency.

      I’d rather just have gpu manufacturers place a cap on how many gpus one can order to the same location each month. Probably wouldn’t work, but I’d rather not pay $300 for a bare minimum 1060 GPU.

      1. Hector says:

        It probably wouldn’t help. The manufacturers aren’t really shipping to individual customers; they usually sell to businesses in bulk. But either way, the problem is that GPU miners are going to grab the units as soon as they become available. They will in some cases call Best Buy (or whatever) every morning just to see if more showed up. Normal customers just can’t compete.

  2. Florian says:

    In fairness, you could argue that being fabless actually increases gpu makers ability to adjust their supply – AMD for example moved some of their production from Global Foundries to Samsung and apparently that allowed them to increase chip production to the point where they are now limited by how much memory they can get their hands on.
    (https://www.anandtech.com/show/12380/amd-to-ramp-up-gpu-production-ram-a-limiting-factor)

    (When talking about mining I usually think of AMD first, since their graphics cards have always traditionally been the preferred choice of the miners since they tend to have better raw computing specs for the same price.)

  3. Florian says:

    One thing that is interesting to note is that most of the demand for graphics cards doesn’t actually come from Bitcoin miners – most of that market has been taken over by custom ASICs years ago.

    I suspect that a large portion of the demand for gpus comes from the second biggest cryptocurrency on the market – Ethereum.

    If that is true then that is potentially good news for the graphics card market since Ethereum is, eventually, going to transition from proof of work to proof of stake, so that might lead to a relaxation in the gpu demand for crypto mining.

    (Although there are enough cryptocurrencies that most miners can probably just transition to one of the other ones.)

    1. Echo Tango says:

      Won’t swapping to one cryptocurrency after another stop being valuable after a while? I mean, this is effectively like buying stock in tech companies[1], hoping for them to go up in value. Except all of the cryptocurrencies are competing with each other, which means there’s probably not much room for more than a handful of them to exist, at the most. How much more bubble / speculation can happen, after one or two of these cryptocurrencies collapse?

      [1] I don’t know the 100% technically correct terms. This is “buying stock” like in the movies. You own a piece of the company, and get dividends, or sell the stocks at a profit.

      1. Florian says:

        After a while, perhaps, but the three biggest cryptocurrencies each have their own justifications for their existence – Bitcoin has the momentum and widespread recognition of being the first, Ethereum has a built in contract system and, if proof of stake goes through, the advantage of not burning electricity like there is no tomorrow, and Monero has certain privacy features that make it particularly attractive to some.

        Of those three, Monero seems unlikely to move away from proof of work that can be done on consumer hardware. (It is, in fact, so optimized for consumer hardware that is somewhat viable, although I doubt, with the inefficiencies of javascript, that this makes much profit, serving instead mostly as a more sneaky way to transfer peoples money to the website owner.)

        1. ThaneofFife says:

          Monero is actually the 10th-largest cryptocurrency right now. The top five are Bitcoin, Ethereum, Ripple, Bitcoin Cash (a BTC fork), and Litecoin.

          Source: https://coinmarketcap.com/

          1. Florian says:

            Thank you for the correction. I guess my brain saw Monero as more important because I’d heard more about it.

        2. Echo Tango says:

          So what you’re saying is that we’ve got crypto miners pushing up the prices for the rest of us, for another decade? Dang; I too, like Shamus, cannot afford a new gaming PC at these current prices. :|

      2. Steve C says:

        Won’t swapping to one cryptocurrency after another stop being valuable after a while?

        No. It shouldn’t. For the same reason that every country can print a currency. Including a separate currency for basically every island in the Caribbean. Some are more stable than others, some grow at different rates, etc. What happens is a complex set of money markets between all of them.

        Note that it would impossible for there to just be one crytpocurrency over a long period of time. The software architecture couldn’t support it on the scale of something like the Euro. Bitcoin itself had to fork already. It became Bitcoin Cash (BCH) and Bitcoin Gold (BTG).

    2. newplan says:

      Really want to second this.

      It’s been years since you could mine Bitcoin with GPUs – pools don’t even let you join with GPU miners.

      Having it in the article is glaring error.

      1. Richard says:

        “Bitcoin” is just shorthand for “cryptocurrencies”.

        Like hoovering your carpet using a Dyson vacuum cleaner.

  4. Hal says:

    So, maybe someone can answer the question of “Is it worth it to dabble in cryptocurrencies?”

    I have a gaming PC that just sits at home for most of the day. Doesn’t do anything while I’m at work or while I sleep. Presumably, I could put it to work “mining” and actually get something out of it.

    Would that even be noticeable income? Would I be worse off because the stress would burn out my PC? I don’t even know what I don’t know.

    1. Droid says:

      Mining at this point is like playing the lottery: You could get lucky and get a 0000000… hash really quickly, but chances are you’re gonna get one right only after a very, very long time. There’s quite a bit of power consumption increase, but the main problem is that your graphics card, and maybe a few other components depending on your setup, are going to experience a lot of additional wear, and that makes it just that much more likely that you will need to buy a new graphics card sooner rather than later.

      For probably around $1000.

      So definitely include that in your calculations if you decide to start mining.

    2. Florian says:

      It looks like, depending on the price of electricity in your region and what hardware you have (I am assuming you have a single card), that you could expect somewhere in the vicinity of a dollar a day if you mined 24/7.

      I don’t particularly think that graphics cards wearing out prematurely is too big a concern since you are generally supposed to undervolt them to maximize performance per Watt and thus profits. (Though that might be too much work to set up multiple times a day.)

      (Note that I haven’t actually tried any of this, and don’t know how one would go about setting it up, either.)

      1. Cilvre says:

        I have two cards in my pc mining on the side, even when i am gaming, and i earn about $100+ a month. It covers their electricity costs and covered buying the second card during Christmas when the costs were relaxed for a bit. Depending on the card you have it can be decent use of it for extra money if you arent worried about the wear. My main card is liquid cooled so I’m not as worried about it.

        1. Steve C says:

          I take it that is your share of a contribution to a pool. Given that bitcoin is ~$10,000 for one nonce.

          1. stratigo says:

            He very likely isn’t mining bitcoin

  5. ElementalAlchemist says:

    You talk a lot about nVidia, but it is worth pointing out that miners are not using Geforce cards by choice, they are using them because all the AMD cards are sold out and that’s all that is left.

    It’s worth looking at the financial statements of both companies for some insight into how mining is affecting sales. The TL:DR is that while card sales have seen a boost, neither nVidia or AMD are the ones really benefiting out of it. All the markup has been on the retail side, so it’s your Amazons and Neweggs that are raking in the real profit right now. Although I’m sure both nVidia and AMD have some ideas in that regard.

  6. default_ex says:

    We have pretty much brought this on ourselves. The GPGPU movement only had one predictable outcome from the start. GPUs would be useful for more than just rendering video. It occurred while usage of computers for playing video games was in a notable growth stage. Compound that with just two major GPU companies producing models that people wanted. How could we not see this coming. One can blame it on cryptocurrencies all they want but it was an eventual result regardless of whether cryptocurrencies became a thing or not. Just hoping it doesn’t last so long that it stagnates the game development industry any more than ‘keeping up with the consoles’ thing does.

    1. Retsam says:

      Hindsight is 20/20, though I’m not sure this was really all that inevitable and obvious.

      Yeah, it’s been apparent for ages that GPUs have uses outside of mere graphics rendering, but if it were just, say, the machine learning people (or the AWS people) that were driving up demand, then in the “normal” course of things there would have been a steady increase in demand, and the manufactures could have steadily increased production to keep up with it.

      But it’s only because cryptocurrency suddenly became a metaphorical Gold Rush that we’re at a “crisis” right now.

      1. default_ex says:

        It was obvious enough I remember reading articles stating much the same back when GPGPU was trying to bend Shader Model 2.0 and 3.0 to computational work. Honestly back then I didn’t believe either way, it was just one of the possibilities as well as the possibility you mention. However I do think it was an eventuality. GPUs, especially modern GPUs rely on ridiculously high quality chips. Not just transistors but also capacitors and diodes need to be of sufficiently high speed to keep those chips fed. GPUs are particularly dense boards when you take the time to look at circuit boards without all the dressings like heat shields, shrouds and plastic covers. Eventually somewhere in the chain something would have given.

  7. Tim G. says:

    Another factor driving GPU consumption is the popularization of machine learning libraries that use the graphics card to power the training cycle so much, much, much faster than the CPU. This means not only individual programmers (and wannabe’s) who wouldn’t have cared much about GPU in the past are now trying to get the latest and great, but also that companies like Amazon that provide rentable computing power are installing thousands of racks of multi-GPU machines available for a few bucks and hour.

    This is less likely to be a bubble than crypto-currency, but it contributes to the current shortage.

    1. Droid says:

      Don’t all the professional machine learning groups use TPUs, or is that just Google?

      1. Tim G. says:

        Last I heard it was just google. I don’t know if they’re even offering it for outside sales yet.

        Certainly might change as the AI market grows and warrants more development for dedicated hardware designed solely for it.

        Amazon is still sucking up nvidia chips for their farms. I’ve heard that they’re working on designing their own AI chips, but I’m not sure if that’s just for consumer devices or also for training servers.

    2. Orophor says:

      Actually beyond both crypto currency and deep learning algorithms the main causes of the shortage are scarce RAM due to smartphones (1.5 billion sold 2017) moving to GDDR5 for fast speeds and lower power consumption.

      At this point the new AMD Raven Ridge APUs with Ryzen cores and Vega graphics are the only sane option if you don’t want to pay at least double MSRP for gaming GPUs.

      1. Fade2Gray says:

        Those APUs make me even sadder. Their benchmarks are getting close to what my aging dedicate gaming PC can push out.

  8. Redrock says:

    The way I see it, there are two ways the whole crypto craze can go. Either cryptocurrencies go back to being a niche instrument with a stable, and therefore unattractive to speculators price, or cryptocurrencies get big enough to warrant massive and internationally coordinated government crackdowns and regulation, which would rather defeat the purpose and, again, discourage speculation. The fact is, the powers that be will not tolerate any real threats or alternatives to fiat currencies and the stock market. You don’t have to call it a bubble, but the current frenzy will die inevitably. The only question is the degree of collateral damage.

    1. Decius says:

      There’s also the situation where an organization, consortium, or emergent behavior emerges where Bitcoin can be converted to currency as easily as currency can be converted to other currency, at rates that are only that volatile.

      The exchange rate of BTC to Dollars and BTC to Euros cannot be more stable than the exchange rate of Dollars to Euros.

  9. ThaneofFife says:

    Shamus,
    I thought people were using ASICs (i.e., specially-designed, single-purpose mining chips) to mine SHA256 cryptocurrencies (e.g., Bitcoin), and were using GPUs to mine Ethereum and Scrypt-based cryptocurrencies (e.g., Litecoin & Dogecoin), as well as newcomers like ZCash & ZClassic (the latter of which has crashed since they switched over to Bitcoin Private).

    Are people still using GPUs to mine BTC on a large scale? If so, that’s massively inefficient compared to the profits they could be making mining ETH and other cryptos.

    1. newplan says:

      ASICs are used for Scrypt coins as well.

      1. ThaneofFife says:

        True, but that’s more recent than BTC ASICs, and I didn’t want to get too far into the weeds. :-)

    2. Echo Tango says:

      Why aren’t the other currencies also using custom chips, if they’re so much more efficient?

      1. ThaneofFife says:

        I’m a crypto noob, but my understanding is that many newer currencies use hash algorithms designed to be ASIC-resistant. For example, I’ve seen Ethereum mining described as something more akin to acting as a single node in a supercomputer than purely running the same hash over and over. I have no idea how that works beyond what I just said, though.

      2. newplan says:

        The short version is that to make your currency ASIC-proof you make the algorithm memory intensive. Doesn’t matter how fast your processing is if it’s waiting for data from memory.

    3. Blake says:

      From the top of the article:

      (Note that when I say “bitcoin” I’m sort of doing this handwave-y inclusion of crypto currency in general. I’m talking about bitcoin because it’s the most recognizable and notorious of the currencies, but a lot of the things I talk about below apply more to the other currencies than to bitcoin specifically.)

      1. Shamus says:

        Whoops. That bit was added after the post went live. Should have been marked with an “edited” notification to avoid exactly this sort of confusion. My bad.

  10. Paul Spooner says:

    I was discussing your previous article with a friend the other day, and hit upon a bright side to this whole scenario, which is that it seems to be incentivizing graphics card producers to design graphics cards with more computational power (which is useful to crypto) instead of more flashy drivers (which is useless for crypto). Maybe by the time the tidal-wave recedes, we’ll see a lull in the graphics card driver conflict as well!

    1. Echo Tango says:

      Could you elaborate on this graphics driver stuff? I thought those just existed to let your OS talk to your GPU; What could be “flashy” / a selling-point here? Also, what is this driver conflict, and how would it be improved / changed?

      1. Paul Spooner says:

        Shamus has done it better, but I’ll try to summarize.

        Graphics card manufacturers all have to use current architectures (since they don’t own their own fabs, see the OP). This means they are all offering fundamentally the same product. So in order to pull ahead of the competition, they often offer fancy game-specific updates to their drivers. In theory these special drivers make the game run better on that manufacturer’s hardware, and has no downsides, leading to a market edge for that manufacturer. In practice all this special code bloats drivers, causes unreliability, and makes troubleshooting a nightmare.

        Because this technique is (apparently?) the only viable strategy for top-end manufacturers, they all do it, which causes interface divergence and endless headaches for developers. The crypto market is offering incentives for bare-bones computational power, which will hopefully lead to reduced-instruction-set graphics cards which will be easier to develop on and less encumbered by legacy driver nonsense.

    2. Daemian Lucifer says:

      That was basically my thought while reading this.Why havent card manufacturers already made special cards that dont include all the graphics driver stuff and are instead geared towards just number crunching?Since the size of the chips cant be changed easily,this would increase their performance for this specific task a lot,thus making them much more desired for mining.

      1. Shamus says:

        About six months ago I read that this was happening / being planned. I’m assuming it takes time to spin up the manufacturing.

        1. Daemian Lucifer says:

          That could drive the prices of graphics cards down,if the supply and advantages of those specialized cards are good enough.Though one problem with those is that you cant really sell them to anyone if mining stops becoming profitable.

          1. Orophor says:

            Even if mining cards were a thing they would not be first choice even for miners. The reason why is their model is built with rapid return on investment (ROI) and frequent upgrades, reselling the older less efficient cards. Other crypto miners don’t want last gen cards either so the resale is typically to gamers.

            1. Daemian Lucifer says:

              That’s because with non dedicated cards you can upgrade to a new generation cheaply by reselling old ones. But a dedicated mining card would be far more efficient than several generations of regular cards, thus you would have no need to upgrade.

              1. Orophor says:

                Alas no, when computing hashes on the block chain it is a race, so if you can get more done than the rest of the pool, you earn more. Thus there will always be demand for faster hashing keeping supply constrained. Like I said in my first comment though the real killer right now is not the crypto mining so much as the billions of smartphones and the limited RAM availability.

  11. The dot-com bubble wasn’t driven by speculators so much as artificially low interest rates. Yes the particular *form* that the bubble took was created by the fad for everything online. But speculation fads without manipulations of the money supply result in skyrocketing interest rates which puts the kibosh on further lending (investment/expansion) until firms start to turn a profit. Firms that aren’t showing a profit can’t afford to borrow more due to the high interest rate, so they are quickly dissolved and liquidated and the resources they were using are put toward profit-making ventures. So, yes, many companies may go out of business due to a fad, but they do it much earlier before they’ve consumed enough resources to put the entire financial system on very shaky footing.

    When speculators can continue to borrow and borrow and borrow, that fuels a bubble. The financial sector becomes heavily leveraged because they’ve lent out more money than they have assets to cover. So when the malinvestments start to fail, it doesn’t just take down the overly-optimistic company that borrowed the money–it threatens to take down the bank that lent the money.

    When it’s a central bank forcing the interest rate low, it often winds up threatening to take down every major bank in the country.

    Speculation fads will always be a thing. A central bank forcing interest rates artificially low does NOT need to be a thing.

    Since bitcoin doesn’t involve borrowing, even if it really does collapse that’s not the same thing as an investment bubble. It won’t wreck banks. Quite a few people might lose their personal investments, but it won’t have large effects on the rest of the economy.

    I think the smart move for Nvidia and AMD and whoever else makes the graphics cards would be to invest in expanded production for their newest cards only. If bitcoin continues to expand, they’re likely to start selling more of the high-price “enthusiast” cards because those cards will enjoy a comparative price advantage for gamers. When last year’s card is way overpriced, why not get the latest and greatest? And Nvidia will then be in position to keep making lots of the newer cards when THEY become last year’s card. That way, they take advantage of the bitcoin situation if it still exists, but they’re still somewhat conservative in not ramping up immediate production too much.

    1. Steve C says:

      That’s not the type of financial bubble being talked about here. A borrowing bubble isn’t the only type of bubble. It’s more of a general bubble. Like the tulips bubble or beanie-baby bubble.

      Nvidia/AMD are in the same place as Ty was (the manufacturer who makes beanie babies). Ty could have greatly increased beanie baby manufacturing capacity to meet the demand of the fad. If they did it wrong though, and their increased supply came in at the the tail end of the fad then they would have screwed themselves. It’s the same decision GPU manufactures are facing right now. That’s what Shamus is talking about.

  12. Steve C says:

    Shamus you’ve focused your articles on cryptocurrency- which is fine. That’s just one part of it though. The technology making cryptocurrency possible is blockchains as you know. Even if all cryptocurrencies everywhere died in a massive bubble, that wouldn’t affect blockchains as a concept. They are a tech that is here to stay. Focusing solely on cryptocurrency is like saying if Google went under then http wouldn’t be used as much. Which would absolutely be true in the short term, but the world wide web isn’t going anywhere even if Google blew up.

    A better analogy is peer-to-peer software. Bitcoin is Napster. Cryptocurrency is music sharing. Blockchains are peer-to-peer. Yes there will be a short term glut of computer hardware if the bubble bursts just as there was in the dot com bubble. The more stable result will be a sorting out the winning apps from the losers, just as it did in the dot com bubble. Napster is long dead. Peer-to-peer file transfers are more numerous now than they ever were when Napster existed.

    There’s more than just cryptocurrency that will be eating up GPU demand by blockchains in the future.

  13. krellen says:

    The whole inflation-spiral things has one central problem: it doesn’t actually happen with fiat currencies that way. The only way US debt will ever become an issue is if the world decides the US government isn’t good for its debts, and even with the credit problems we had under Bush II, US Treasury Bills are still just about the best investment on the planet, because the US always pays its bills.

    Short of global economic collapse, concerns about US government debt are largely misplaced.

    Now, other countries may not be so lucky, but it’s my impression that crypto is a largely US-driven thing.

    1. Jeysie says:

      Yeah, this.

      One of the reasons Bitcoin is such a bugaboo to address effectively directly is that a lot of motivations behind it are economically unsound, morally dodgy, or economically unsound AND morally dodgy.

      And those reasons to use Bitcoin that are fairly sound and reasonable motivate only a very small fraction of Bitcoin users and would probably be much better served by a currency option that isn’t devouring computing time and electricity like a technological elder god.

  14. Daemian Lucifer says:

    Another big hurdle for making more cards is the nature of quantum.Seeing how modern semiconductors are just few atoms thick,being precise with their manufacturing becomes increasingly difficult.The difference between a good and great graphics card(any chip,really)has become less dependent on what was cheaper to make and more dependent on how often the perfect configuration would occur.

  15. Zaxares says:

    Yeah, I think a lot of bitcoin miners bought expensive cards thinking that they’ll be able to resell them afterwards and further recoup their initial expenses. But the market for used graphics cards is pretty small, since PC gamers like myself tend to shy away from used cards (for all the reasons you mentioned) and go for more top of the line stuff. That means that buyers for such a market are going to be people who don’t need powerful cards OR can’t afford the good stuff, both of which translate into “poor resale prices”.

    As for bitcoin itself, I think that the cryptocurrencies themselves will eventually settle down or die out. They are essentially a store of value, and eventually people will realize that a store of value that has no inherent value in itself (like gold, because people always have and always will like having pretty, shiny stuff to make jewelry and ornaments) AND isn’t backed by any kind of central authority (which cryptocurrencies lack) is one that is highly unstable and prone to a crash as soon as people stop believing in its value. That said, I think the technology behind cryptocurrency as far as securing transaction integrity WILL stay and eventually be adopted by the mainstream financial sector.

  16. Preciousgollum says:

    Ever since I built my first and only custom PC back in 2006-7, I’ve seen it as a bad investment. 7900gt for £250 of lucky,

    Because ‘The PC’ is no longer the only option for the average person seeking computing power (we have phones, tablets etc), custom PC building has been on the decline for the past 10 years, and has been reduced to a hobby. Every time I wanted to build a new PC, i’d find new ways to side-grade with second hand components, or even buy a full second-hand machine from those that were selling (because they had made a bad investment as an early adopter) – or I’d realise that I didn’t really need it in the first place. Many a custom build wishlist has gone unpurchased.

    It happens with anything new – HDTV or Blu-Ray/HD-DVD or Xbox 360 vs PS3. Eventually I’ve precured them all at reasonable value, and/or they become obsolete anyway.

    A GTX 1070 might be huge amounts of money for a ‘4K ready’ gaming PC, but you can buy an OEM GTX 960 4gb for just over £40 direct from China. You just have to wait for 2 weeks to a month for shipping.

    So yea, early adopters might suffer, but everyone else has it fairly ok. Perhaps the demand for adopting new horsepower tech just isn’t as great as people want it to be.

      1. Preciousgollum says:

        Thanks for the warning on ebay GPU.

        Also, I realise I was in the middle of writing, and then forgot and moved onto a different one lol. It was about how any top of the line gpu in 2006/7 was unreasonable (7900GTX 512MB) and that decent GPUs have always been ‘lavish’ in their expense – often double the amount of a Console.

        What I’m saying is that it is entirely possible the consoles are having an effect on the market – Sony, Microsoft, AMD + Nvidia etc don’t exist in a vacuum – perhaps they are all trying to make their respective closed platforms more appealing by making open platforms less viable. It has at least felt like this to me for about the past 8 years. This time, the reduction of the open platform (PC) can be blamed on another open platform (Bitcoin Mining) which then tarnishes the value of both quite effectively. Thus, people will flocked to closed platforms, in their ire of the open ones.

        Would people have had the same reaction if GPUs were sold out for something philanthropic like Folding@home?

        Other ‘first party’ or ‘official’ products have been creeping up in price, such as Xbox 360 AND Xbox One controllers + they are also breaking too often.
        Any Arcade Stick that is of decent quality also cost quite a sum.

        In summary, I don’t think that bitcoin mining is the only reason why GPUs cost more. I think there are other reasons, although I couldn’t articulate all the possibilities very well.

    1. Zekiel says:

      I was a faithful PC gamer for almost 20 years. Switched to PS4 last year because the comparative prices to run the same AAA games were ridiculous. It was like £750 to buy a PC capable of running Witcher 3 at recommended specs. £220 for a PS4.

  17. tankermottind says:

    I think for future consumer level cards, nVidia should just lock out their GPGPU features and make them useless for crypto mining, or any purpose except 3D graphics. If you want to mine coins you’ll have to get a Quadro.

    1. Droid says:

      That screws over everyone, even the gamers, for no real short-term benefit in price.

      As much as I hate the current situation, mathematicians, physicists and probably a ton of other scientists rely on GPGPU programs on their home computer both during their college years and for any side projects they might be working on, in addition to those people who work from home.

      And it’s not like games that aren’t the biggest graphical gems, but nonetheless require heavy number-crunching every frame wouldn’t exist (*cough* Factorio, *cough* Paradox games, *cough* AI War *cough*).

      Sorry, could anyone spare a mint or something? *cough*

    2. Richard says:

      PhysX is GPGPU

      Cloth, hair, lighting – all of that is GPGPU.

      At work I do GPGPU using OpenGL ES 2.0. All you need for GPGPU is a programmable pipeline – shaders – and a way to get the results back.

      Even if you have to use glReadPixels() to get the results, it’s still worth it for a lot of tasks.

      The usual limitation is that floating point maths isn’t fully compliant with IEEE 754.
      However, that doesn’t really matter as it’s the weird edge cases that are noncompliant, and it’s easy enough to characterize the specific hardware you’ve got and work around them.

  18. poiumty says:

    I get GPUs, but what’s with memory prices though? A twin stick of dual-channel 16GB DDR4 (total) costs about as much as an i5-8600. That… doesn’t seem right.

    1. Orophor says:

      The same root cause, demand for RAM is outstripping production capacity, which makes the price rise. There are only four major memory chip manufacturers: Samsung, Toshiba, SK Hynix, and Micron. There are a lot of new demands: smartphones, cloud scale data centers, cryptominers, shift to DDR4 for both Intel and AMD platforms, and perhaps some price fixing: https://www.gamersnexus.net/industry/3186-chinese-government-investigating-dram-price-fixing-allegations

  19. Philadelphus says:

    Another way to think about nanometers is that most atoms have diameters on the order of a tenth of a nanometer (or an Ångstrom, as we call it in astronomy).

  20. Humanoid says:

    Maybe all the big game devs should get together and announce that all of their big upcoming titles will be in 2D. The next Battlefield, CoD, AC, GTA, all in 2D. Fallout 5 will be done with the Battlespire engine.

  21. Fade2Gray says:

    I’m definitely at the “praying my card doesn’t die before prices drop” phase. Aside from the usual signs that it’s time for an upgrade (dropping presets to medium on anything released in the past year to reach something close to a stable 30FPS) I’ve started to notice certain graphical glitches consistently cropping up across multiple games which has me worried that something might be giving out on the card itself.

    So, umm, anyone need any organs? I think I have a few I could live without. 33 years typical use in a non-smoking home. Willing to consider trade for modern GPU.

  22. Blackbird71 says:

    Back in 2007, circuits were around 100 nanometers[2]. Every few years we refine the process and the circuits get a little smaller. 40nm. Then 28nm. 20nm. 14nm. Last year, TSMC began selling their first 10nm chips.

    One correction/clarification: circuits/chips built on a 10nm process are not 10nm in size. 10nm is the size of the smallest feature that can be made on the circuit. The circuit itself is only limited by the amount of silicon you want to use to build it.

    Think of it like designing the roadways of a city. If the smallest path you can design will be a single lane road, then you can say that you are designing on a “one car width” process. However, your actual roadways will span the whole city, which can take up however much real estate you have available. The more land you use, the more expensive it is, so of course you will try to design the smallest and most efficient layout possible that still meets all the needs of your city plan. Whatever your design, the total road system will be much larger than just “one car width”.

  23. Mousazz says:

    Wait, you MADE the memebux image? Wow. I’m genuinely impressed.

    1. Droid says:

      Yeah, I was too, but forgot to comment.

      1. Daemian Lucifer says:

        Same with me.Only I forgot to even reply to Mousazz.

  24. Joe says:

    There is some debate on the issue, and I don’t claim to speak for the bitcoin-mining community (don’t even mine it myself) but from some preliminary research online it seems most miners tend to under clock their GPU’s, as it’s more energy-efficient and the cards last longer, meaning less instances of re-buying any that fail from being overworked.

    As miners are interested in their bottom line most of all, it seems it’s in their best interests to keep the GPU’s as healthy as possible for as long as possible, so it seems buying a second-hand card used for mining might be a good medium-to-long term investment, depending on the miner you bought it from.

    We’ll have to see what happens after the cryptocurrancy bubble bursts (if ever), you may find some sellers of GPU’s have been treating the cards well and are perfectly serviceable for gaming. Others may have not cared or were ill-informed about hardware maintenance. Guess the advice is the same as it’s always been – do your research on the seller, be careful who you (metaphorically) jump into bed with.

  25. Mr. Wolf says:

    “Yeah, I used this image last week. But I foolishly spent like two hours making this silly thing and now I’m trying to justify burning those hours on a two-second sight gag.”

    By my calculations, you only need 3600 people to see it to break even.

Thanks for joining the discussion. Be nice, don't post angry, and enjoy yourself. This is supposed to be fun. Your email address will not be published. Required fields are marked*

You can enclose spoilers in <strike> tags like so:
<strike>Darth Vader is Luke's father!</strike>

You can make things italics like this:
Can you imagine having Darth Vader as your <i>father</i>?

You can make things bold like this:
I'm <b>very</b> glad Darth Vader isn't my father.

You can make links like this:
I'm reading about <a href="http://en.wikipedia.org/wiki/Darth_Vader">Darth Vader</a> on Wikipedia!

You can quote someone like this:
Darth Vader said <blockquote>Luke, I am your father.</blockquote>

Leave a Reply to Orophor Cancel reply

Your email address will not be published.