Sure if they charge themselves the same $5/TB that they charge customers. The only question is can they plot efficiently. I fail to see how they couldn’t farm/harvest profitablity. The article is lacking any hard math past “we can’t do it”, if I remember correctly, it’s been a few days since o read that one.
Also their statement about netspace growth rate it outdated.
A couple things. ANYONE thinking they can keep up with netspace without access to endless super cheap/free storage is going to be fighting a losing battle. That doesn’t mean chia is unprofitable to mine. it just means it will have ever tightening margins that must be watched. If backblaze plotted in ram with madmax there would be zero SSD cost, in fact zero wear cost at all, just electricity and the plotting hardware investment that can easily be reallocated.
Chia seemed very much like a data center crypto when I first ignored it for that reason. It is ideal for data centers to offset costs, but only if it’s plotted efficiently and used on over provisioned space that can easily be deleted when it is needed. The thing is that home users actually do have a chance of undercutting data center costs, because chia doesn’t require paying for power or data redundancy since chia plots are not critical data.
People need to be realistic that chia farming will absolutely have tighter and tighter margins as netspace grows, especially as pools average out rewards even more. Chia farming profits will absolutely come down to the cost of electricity at some point for many farmers. None of that is a bad thing for chia crypto because it will take huge decentralization for power cost to become the deciding factor for profits.
No one should be buying storage that they have no use for beyond chia. Data centers will have to do ongoing cost/profit monitoring to decide when they are no longer making profits from it. I personally used chia as a reason to upgrade my home plex server capacity with affordable used enterprise storage. I won’t be upgrading that for a long time and now 4k video doesn’t seem so crazy
A couple things. ANYONE thinking they can keep up with netspace without access to endless super cheap/free storage is going to be fighting a losing battle.
I still can’t understand the difference of Chia from other crypto like ETH for this statement, would it be more difficult also for GPU mining that you earn less and less ETH per MH/s as more miners join the game?
One thing I agree is, you could spin up and down a GPU almost instantly, while for Chia you cannot spin up a farm; you will have to plot. That might suggest an “inertia” in Chia’s difficulty comparing to others, but in terms of keeping up with netspace (difficulty in general), wouldn’t it be the same as everything else?
Yes, it is the same as any proof of work, as the difficulty raises your returns drop. People chasing bitcoin profits is how we ended up with almost all the coins already being mined and warehouses of asics blasting through electricity. The thing is Eth mining profits now are from transaction fees, not just the coin it mines. Chia hasn’t even started having transaction fees since it’s just a blockchain at this point, the real work of making real world value from the blockchain being used is what’s coming. I’d be surprised if XCH transaction fees amount to anything until a farming rate halving or even two. The plots made today should be just as usable then as now.
Very interesting, I am new to the game and actually just started also mining ETH (with my multiple idle Vega and RVII). I learnt that in ETH they are kind of reversing the course where there will be EIP1559 on Aug-4th, essentially reducing the mining profit from transaction fee.
I wonder how would this affect the direction other cryptos might take in rewarding miners/farmers, which if I understand correctly from you is an important part of the profit.
It should increase the price due the deflationary implementation of EIP-1559. However, the bigger “problem” is the change to Eth 2.0 but it will take more than a year for sure and you can recover the investment in that time and then sell the gpu
Not counting the fact that with such a high rate of hard forks they can fuck up the chain at anytime in the process of 2.0 migration… For me that’s quite a risk, and they should be careful and take as time as they need to make it right.
This is not true, until you’re pushing everything into RAM. But following recommended settings, it is not true. It’s still using SSD for saving and compression of plot. It is around 25% of standard plotter disk usage. 1/4 of standard usage still is a lot for consumer SSDs.
But actually there is Bladebit, but on non-commercial licence.
If your RAM disk is large enough, which is trivial for an enterprise, you can still run madmax entirely on ram I believe though with marginal speed up.