Came across some curious figures on CHIA mining (acutal for May 5-6), see for yourself what it turns out:
After yesterday’s conversations with colleagues about SSD mining, I decided to play around a bit with the new CHIA coin (https://www.chia.net/).
In 2018 I already addressed this topic, but instead of ROI there was a monkey business, so a regular HDD would take several years to pay off. So I wasn’t interested in these fabulous profits then, and I extrapolated the same conclusions to new cryptocurrencies, until CHIA was worth $2,000. Now it’s down to $600, but even that’s a pretty interesting price, so I decided to see what a server from Hetzner could bring.
Servers and setup. Beginning of the experiment
It should be noted that servers in Hetzner have gone up in price by about 10 euros in a month, so the 1700 costs 73 euros instead of 63, for which I bought a server for some of my projects in April. And before the new year, servers were generally 55 euros. EPYC AMD prices start from 260 euros, although a few months ago they were 120-150 euros.
Setting up is elementary, like everything started and even ran a complete synchronization for an hour, but after that 15 thousand blocks turned into 240 thousand, so, apparently, will be synchronized before noon.
During the hour of synchronization the difficulty of mining increased from 7 to 8, as people are actively entering the topic with CHIA mining along the way. Two plots (hard disk space) do somewhere around 8% per hour, which means that the plot will roll out in 7.5 hours if everything goes the same way. But it’s not clear if it will work in the absence of full synchronization.
So tomorrow or the day after tomorrow there will be a day of revelation, when it becomes clear that the server build for 120 euros (if we finish with a 2Tb NVMe SSD for plots and a 4Tb HDD for storage and mining) will pay for a year.
Part Two. Expectation vs Reality
Something about the fantasy of finishing work at noon (see above) turned out to be a pipe dream and synchronization of the network, apparently, will take the whole next day. Given that the currency is new and has existed since March, at this rate in a month to synchronize will have to weeks, while the Digibytes I have synchronized approx. an a day for about three years.
But there is hope that they will improve the performance here as well, because I only killed about an hour to get this very synchronization running.
Remarkably, the difficulty from 7, last night, went up to 19, and as I write this post, became 20:
Apparently it is not possible to get rich, because the rate continues to fall and I suspect that in theory it should go below the payback on the servers.
Just finished CHIA synchronization, which took almost 20 hours.
It’s funny, because in the morning the capacity of CHIA pool was 270 petabytes (10 to the 15th degree) and the complexity was 20, now the capacity is 2.5 exabytes (10*18) and the complexity is 170 – so people are involved in mining the whole day, increasing the capacity for 12 hours, which of course affects the price of coins.
Well, three plots of 300 gigabytes, which were formed during the synchronization, judging by the table, will be filled with CHIA in five years, that is, the server in principle will never pay off.
If you rent a 16TiB hard drive, it will of course increase the size of the plot by 50 times, but judging by the increase – it is a pointless race, because by the time you organize a pool of 16TiB, X will become Z. Considering that even now 16Tb from 3Eb is 0.00001%, that is just as much over the limits as 300 gigabytes. And we need more than 300Tb.
CHIA mining – server rentals will pay for themselves in 5 years.
On that solemn note the experiment can be considered finished, so the extrapolation from three years ago was absolutely correct.