It is not correct to apply that to SAFE being compared to BTC.
And apparently the power usage of a global SAFE is not even going to be that of BTC now, let alone BTC in a few years. Obviouly ignoring clients since that power is already being used for browsers.
Then additionally SAFE vs BTC is not even in the same technological neighbourhood which the paradox is based on. In other words SAFE is not a technological advance on BTC. A tiny portion of SAFE’s operation is an advance on BTC but 99% of SAFE is nothing to do with advance on BTC technology.
Its like saying that 100 story building is a technological advance on the truck because in the building is a backup generator. Its apples and oranges
Assuming 25% market penetration for single user vaults gives us ~1 Billion, and an average power differential for a single vault owner of 10 watts (low end for CPU/MOBO/RAM+SSD/HDD, could be a lot higher) puts an established global safe network at 10GW = 87.60 TWh per year.
For comparison there is a wide variance in bitcoin energy consumption estimates. I think a lot of them are over-hyped (note: I am not a fan of BTC). These articles put some aspects in perspective without too much msm hype.
From a psychological perspective, I think that the lightbulb test comes into play. Most individuals won’t care how much differential power they consume running a safe vault 24/7 until it surpasses the equivalent of one 60watt lightbulb (with some/all costs somewhat offset by safecoin revenue).
So for a very simple “back of the envelope” estimate, I don’t think its unfair to say that on average SAFE farming will consume approximately 60watts x number of vault owners.
Yes, SAFE and SafeCoin will offer orders of magnitude more utility than any other crypto ( … or all others combined) but it’s still going to use a lot of energy. Which is ok, because that’s progress, solar panels are cheap now, and fusion is only 20 years away.
10w sounds about right for dedicated devices, e.g. a 4tb Western Digital MyCloud device uses around 9w when active, and wouldn’t use much more with a slightly beefed up processor for all the crypto work a vault would need to do.
If people already have computers / NAS / routers running that could be vaults, for those computers, the additional power consumption of running a vault will be a fraction of 10w.
That ~90TWh of energy may sound like a lot, but that’ll probably be countered by a reduction in demand from data centres.
True… in 20 years time Solar & solid state storage will be so cheap people will wonder why anyone thought achieving clean energy would be a problem… and fusion may still be 20 years out of reach
Good point/ nice find. This is an interesting figure that I didn’t take the time to look in to. But you are right on. Taken to the limit of global adoption the safe network will consume more energy than all the world’s data centers combined (because they will all be SAFE). However, deduplication will and automated redundancy will also be a huge efficiency gain in this realm. I still don’t think its unfair to say that a future global SAFE network will consume at least as much as all current datacenters combined, since technology/efficiency improvements will handle the orders of magnitude increase in data for similar energy consumption to what we have currently.
Highly unlikely. We are talking of concurrently running vaults. Considering over half the world is unlikely to ever have a desktop or similar we have to assume they will use a mobile device if they have a vault. So 8 hours a day maximum for those (on charge).
I estimate that in the first 5 to 10 years we will definitely have less than your 1 billion and more likely 250 to 500 million concurrently on at the same time. Many will be phones/tablets/etc which will only need 1 watt incremental power. So half the world’s population’s vaults will draw < 1 watt
Then incremental power for home users will be in terms of 5 watts
Then power requirements for those who use a data centre will be like 10 watts incremental since they are running instances with other instances running on the same hardware
vault number unit ave on Power
type vaults power time Usage Ave
----- ------ ----- ------- ---------
mobile 500M 1W 33% 167MW on charge
home 500M 5W 80% 2000MW Some 24/7 rest on/off
Data/C 100M 10W 100% 1000MW Assume 24/7 for all
This give a usage of 3167MW (3.2GW)
That is so way off target. You even gave the total use of a harddrive. 60W of incremental power requirements is bogus in my calculations. My laptop does not even draw that power when I use it, and the incremental power for running a vault is <5 Watts max. About a max of 100WHr per day whereas your 60 Watts is 1.44KWHr per day. Even for a desktop its <6 Watts max incremental power. Even for a datacentre where a person rents an instance to run a vault its only <10W since the instance is running concurrently with other instances and disk is only a portion of a Network storage solution.
Today BTC is estimated at 2.5GW now and growing steadily
SAFE at full penetration (in 5-10 years) is est at 3.2GW assuming maximum power usage profiles for incremental usage and maybe 30% less. Also assume high adoption of vault usage.
So I have no problems saying SAFE power will be a lot less than BTC
Earth electrical power usage is at 20.0 TW (20000 GW) in 2016 and higher today. So the additional power usage of SAFE vaults is going to be less than 0.05% and its a global network. Remind me how much the internet infrastructure draws? And what %age will vaults be of that? Now that would be an interesting statistic.
This was an average, while also considering the incremental cost increase many people may be willing to bear to grow and support the network. Consider all the crypto miners now chugging away with multiple 1500 watt psu today. Those same people are going to be fine with running 1500 watts of server/hdd.
Yes, there may be a lot of folks running lower power devices, but there will be a lot of high performance devices too.
Yes, this is another good metric.
70Twh in 2016 says forbes.
Considering they can do this for 10 Watts on a power hungry SBC (< $100) how you can say this to be 60 watts is beyond me. Unless they use their pre 2015 computers to do it and only use the computer for one vault. Oh and never user laptops for a few vaults (20-50watts for laptop and 3+ vaults dedicated)
And at 10 Watts that means they are running 150 vaults of the 1 billion vaults. It doesn’t matter how much power you draw but how much per vault out of the world’s total vaults.
So that is a bad way to look at it. They will be saying, OH I can run 10 vaults for 30-100 watts and save 1400 watts. Great lets do that. And you forgot that running 150 vaults on their internet connection may not work while it did for mining.
so 8 GW
Now how much of that 8 GW is for the servers of the internet?
How much power REDUCTION due to those servers being retired and the Data now on SAFE? How many will become the data centre vaults?
So again I assert that SAFE will actually cause a power reduction when crypto either migrates or reduces due to SAFEcoin, the internet servers being retired (or becoming part of my vault power estimates above). And the basic reason is that data centres will reduce and migrate to spare resources of 1 billion people and instead of powering datacentres we start using power already being used.
Two words, greed and/or altruism. Also consider the case when vaults/nodes are also providing compute resources. Depending on application, people using SAFE for SafeCluster and HPC aren’t going to be happy with the performance of an SBC, tablet, or phone, no matter how many you throw at them (some applications might work fine if the cost is low enough). In that case they are going to want the state of the art equivalent of today’s threadripper and GPU counterpart. (Bonus for cryptonaughts: All the current ETH miners will end up with a new market to keep their gpu rigs going.) This will eat up a lot of power, but it will be productive. Consider the implications of unleashing the TFlops of BTC on scientific research, protein folding, material modeling, fusion simulation, boinc projects, etc.
Hello, I read everything posted on this topic and I really enjoyed it, I must say that the over consumption and misuse of energy on a global scale is destroying our natural reserves and devastating the world we live in today. If not put a brake on this what will be of the future of this planet? not to mention that we are still exploited by operators of electric power supply. In a country where a lot of energy is consumed and the price paid is too high, do you think it is a good idea for every citizen to create his or her own energy generating medium? Using solar panels for example? This would somehow force the distributor to lower the fees charged to consumers which in my view is a theft in our pocket every month. What do you think about this?
With solar power getting cheaper & more efficient all the time, and solid-state battery development moving forward all the time, it’ll soon be normal for solar power & storage to be built into all new buildings & to retrofit to older buildings.
If it’s the most economic option, it’ll happen, and that might only be a few years away.
Add in smart mini-grid technology, and I’m sure the days of the big central energy suppliers are numbered, though I’m sure larger scale operations will still be needed to power high population density cities.
However, I can’t agree that energy companies providing services is theft - it’s an expensive task to produce power & distribute it widely & reliably. Of course, some energy companies will be ripping people off & abusing monopoly power, so in these cases something should be done, but markets for ‘natural monopolies’ are hard to get right… decentralisation of generation & distribution could revolutionise the quality of energy markets even if people don’t generate the power themselves at home.
If you do the math, photovoltaic solar panels are the least expensive form of on site power “generation” right now. The big issue is “storage”. If you live in sunnier climates you can get away with less but it still takes up front investment. Elon is helping to bring us from about $0.20/kwh down to $0.05/kwh life cycle cost just for batteries.
