22

In The Guardian, on May 29, 2021, Adam Greenfield writes:

Elon Musk’s recent large-scale transactions in proof-of-work-based Bitcoin released more carbon into the atmosphere in just a few days than the amount saved, in principle, by all the Teslas ever sold.
("Non-fungible tokens aren’t a harmless digital fad – they’re a disaster for our planet")

Is this possibly true? They've sold somewhere around 1 million Teslas, so this seems like a pretty big claim. I read that bitcoin overall is releasing more carbon than all of New Zealand, but I don't understand how to attribute the specific carbon emissions of a transaction. Or is the key to this sentence that Teslas per se don't reduce carbon very much because the electric grid is still high on carbon?

krubo
  • 724
  • 5
  • 9
  • 2
    I don't really like the _"the amount saved, **in principle**"_. How is that handled? Do they explain something more? – pipe May 30 '21 at 21:26
  • It's hard to know what the exact claim is here. The claim could relate to the transaction somehow, or it could apply to the mining cost of the coins transacted (very much not the same thing, but almost certainly easier to estimate). – Bryan Krause May 30 '21 at 22:40
  • 5
    I understand ”amount saved, in principle, by all the Teslas ever sold” as: all Teslas together have consumed electricity, releasing *x* CO₂. If people had driven gas-powered cars instead (plausible assumption: changing to an electric car didn't fundamentally alter people's driving habits), they would have released *y* CO₂. The amount saved is *y* - *x*. – Gilles 'SO- stop being evil' May 30 '21 at 22:47
  • 4
    This doesn't make sense. It's **mining** Bitcoin that's energy intensive, not transactions with existing Bitcoin. – Loren Pechtel May 31 '21 at 04:12
  • 2
    @Shadur - but presumably every transaction doesn't require more mining - once a bit-coin is mined it can be spent back and forth. Like transactions in paper money don't require cutting down more trees, only printing more money does. – komodosp May 31 '21 at 07:12
  • 10
    @colmde this is an incorrect perception. What is referred to as "mining" is a process of transaction validation. Since this process is a resource-intensive one, a person facilitating it receives a reward - but that, from the cryptocurrency infrastructure viewpoint, is a side effect, not the intended purpose of the activity. You could compare it to pen-and-paper accounting - any transaction requires to be recorded. In a book. Made of trees =) – Danila Smirnov May 31 '21 at 09:29
  • How much was transacted? – LShaver May 31 '21 at 12:05
  • 1
    This is a complex question. If you believe that Elon Musk's Bitcoin transactions increase Bitcoins market cap by 10% then that means that Bitcoin miners would be incentivized to spend 10% more on electricity. It however takes a while till the miners buy more mining equipment and mine more and burn additional energy to make a profit on the increased Bitcoin price. It's however quite unclear whether Elon Musk's transactions had that effect. His action might very well have been net negative for Bitcoin's price and thus reduced Bitcoin's emissions. – Christian May 31 '21 at 16:18
  • 1
    @DanilaSmirnov Resource intensive compared to traditional accounting, but nothing like the scale that is being talked about. The claim isn't obvious nonsense if it's comparing the carbon used to mine the bitcoin, but once a bitcoin is mined it's not that expensive to move it around. – Loren Pechtel Jun 01 '21 at 01:33
  • 2
    @LorenPechtel I'm not going to compare scales of bitcoin mining power consumption to other processes, but you can't separate "moving it around" from mining, because if your transaction is not validated ("mined") - *it did not happen*. Mining as in "creating new coins" can happen without transactions, but the percentage of empty (i.e. containing only the coinbase reward transaction) blocks is decreasing over time. In 2019, only ~0,8% of mined blocks were empty; so I'd say at the moment carbon footprint of pure "mining" is negligible compared to transaction validation "mining". – Danila Smirnov Jun 01 '21 at 03:44
  • 1
    It makes no much sense to compare energy spent on Bitcoin transactions with energy saved by Teslas. The latter value is a differential value (difference between two alternatives: riding Tesla or riding ICE car), while the former value is an absolute value, not a difference between two alternatives. Hence, it becomes a typical "apple vs oranges". A correct method would be to compare two differential values: "energy consumption difference after replacing ICE cars by electric cars" versus "energy consumption difference after replacing traditional payment systems by cryptocurrency". – Bronx Jun 01 '21 at 21:53
  • 1
    @Bronx this isn't an academic treatment trying to compare the two values -- it's a journalistic statement trying to help the audience conceptualize the carbon impacts of cryptocurrency. People generally understand that switching to EVs is important for reducing carbon emissions, and people generally know that bitcoin is energy intensive -- comparing them this way (while not valid, as I explain in my answer) is an attempt to express just how intensive bitcoin is. – LShaver Jun 04 '21 at 15:23

1 Answers1

15

tl;dr: The claim cannot possibly be true, unless "Elon Musk's recent large-scale transactions" refers to all bitcoin transactions over the last six months.

"The amount saved..."

A detailed analysis by Carbon Brief finds that over its lifetime (assuming 150,000 km driven), a Tesla Model 3 results in 10 to 19 tonnes of carbon emissions, depending on the country (and subsequent carbon emissions of the grid).

Comparing models on fueleconomy.gov, the Model 3 is the most efficient vehicle in Tesla's lineup with a current rating of 142 miles per gallon electric (MPGe), compared to 89 for the original 2014 Model S.

Tesla Model 3 vs S efficiency

Taking the Model S as the low end, a Tesla results in 14 to 29 less tonnes of CO2 emissions than a conventional vehicle over its lifetime. This is compared to the average new conventional vehicle sold in Europe in 2019. If the efficiency of all cars sold in all Tesla markets since 2014 were the basis of comparison, the high end of the savings range would likely increase.

"... by all the Teslas ever sold"

Sources on Tesla sales to date include reviews of Tesla quarterly reports and some estimates. This chart from Statista reports 1,474,900 Teslas sold globally from 2016 through March 2021. Another source reports U.S. sales of 42,105 Teslas from 2014 to 2015, bringing the total to 1,517,005. Allowing for some global sales in the early years, and additional sales since March 2021, a figure of 1.6 million Teslas sold to date seems reasonable.

Combining the per-vehicle savings range with the total Tesla sales to date results in savings of 23 to 46 million tonnes CO2 emissions reduced by all Teslas sold to date.

Bitcoin transactions "in just a few days"

According to the digiconomist blog, the carbon footprint of a single bitcoin transaction as of March of this year is 706.47 kg CO2.

In May, there have been approximately 252,000 bitcoin transactions per day, for a total daily emissions of approximately 178,000 tonnes of CO2.

At this rate, emissions resulting from bitcoin transactions will equal the emissions saved by all Teslas ever sold in 129 to 257 days. Put another way, in about 200 days, bitcoin erases all emissions reductions from all Teslas ever sold.

While this is staggering and depressing, it's still significantly long enough that we don't have to figure out what the quote means by "Elon Musk's recent large-scale transactions."

LShaver
  • 9,931
  • 6
  • 41
  • 73
  • 3
    "the carbon footprint of a single bitcoin transaction as of March of this year is 706.47 kg CO2." Whether transactions are made or not, whether they 10x increase or not, is in no way related to the energy spent mining bitcoin, thus it is non-sensical to make a "carbon footprint of a bitcoin transaction" metric. If we did, we would need to clearly recognize that it is arbitrary and necessarily a cherry pick unless the whole of all mining and bitcoin transactions are included in the calculation. –  May 31 '21 at 22:12
  • 3
    @fredsbend that statistic is direct from the linked source, so perhaps subject of a different question? Regardless, my understanding (acquired today) is that **mining bitcoin** and **processing transactions** [cannot actually be separated](https://en.wikipedia.org/wiki/Bitcoin_network#Transactions). The mining is how transactions are verified and distributed to the ledger, so one cannot occur without the other. – LShaver May 31 '21 at 22:34
  • 2
    Transactions cannot occur without the mining, but mining can occur without transactions (and indeed did for many blocks when bitcoin was new, called and 'empty block'). I'm not saying that bitcoin mining doesn't verify the transactions, but I am saying that the cost it takes to mine is not at all connected to the number of transactions. Submitting the block to the network with all the transactions is negligible in cost to the miner. The "proof of work" shows that the miner was willing to spend great effort more than a spoofer would, therefore the block is submitted trustlessly. –  Jun 01 '21 at 00:28
  • 1
    @fredsbend I'm not enough of an expert to really argue the point (I'll lean on the source which was cited in a few other places), but it does seem like saying mining and transactions are "not at all connected" is a bit strong. It's a currency, so if people weren't using it, no one would be mining it. There's **some** kind of connection, even if it's more the "invisible hand" sort. – LShaver Jun 01 '21 at 01:16
  • It's the same nonsense as with the claim that nfts are producing vast amounts of carbon emissions, it's simply not how it works, the carbon is released independently of the creation of nft or Bitcoin transaction. Of course there's some linkage between the two but the direct attribution articles like to make is again, nonsense. – eps Jun 04 '21 at 14:24
  • Basically in both cases they are trying to divide the sunk cost of carbon emissions due to mining over the number of things (nfts, transactions) that occur. If for some reason Bitcoin transactions increased by orders of magnitude the cost of carbon per transaction would go to zero. Of course in reality an increase in transactions would probably mean the price is going up as more people buy in which leads to more mining but that's not how these sorts of articles are written. They always draw a direct link, when really elon's thing had no direct impact on emissions at all. – eps Jun 04 '21 at 14:36
  • 2
    @eps again, I think "nonsense" is an overly strong characterization. When I buy a gallon of gasoline it has a negligible impact on carbon emissions -- it could even result in a reduction, if I'm going to put it in a can and bury it in the ground. But when a million people buy a million gallons, we can compute the emissions impacts, then divide it back to that one gallon to get a better idea of the impacts. That's the same thing that's happening here -- the bulk impacts of bitcoin are being divided across all transactions to get an understanding of the total system impacts. – LShaver Jun 04 '21 at 15:18
  • @Eps we do this sort of thing *all the time* with per capita measures, carbon intensity measures, trade flows, etc. Why is carbon emissions per bitcoin transaction not a valid way to understand the bulk carbon impacts of bitcoin? Or is there a better way to generalize and conceptualize the emissions impact? – LShaver Jun 04 '21 at 15:20
  • @LShaver it's not "invalid" but it is perhaps a bit peculiar if you understand the technology. It's analogous to measuring carbon impact of the space shuttle program by dividing total carbon emissions by number of satellite maintenance procedures done. While satellites were a key source of funding for the shuttle, some shuttle launches had nothing to do with satellites. It would make more sense to divide total emissions by number of launches as a measure of the carbon impact of the space shuttle. – Chan-Ho Suh Jul 17 '21 at 18:47
  • @Chan-HoSuh Your analogy makes sense to me, but what would you propose as a more meaningful metric for bitcoin? – LShaver Jul 17 '21 at 19:23
  • @LShaver Only the total energy spent makes sense. "Bitcoin, as an industry, has spent x energy since arbitrary date." What we get out of that is objective, but whether it's worth it, subjective. –  Dec 31 '21 at 23:00