[ksec]

Some people think Chia is so much better because of it uses SSD / Storage rather than GPU. I think this could be far worse.

Not everyone needs a GPGPU, but everyone need NAND and SSD. And it already causes price hike for SSD we will soon see something similar to chip shortage. The DRAM Fab which could be quickly retooled to NAND will be part of this equation as well.

This is not good.

[marcusverus]

> And it already causes price hike for SSD we will soon see something similar to chip shortage.

Nah. The expected ROI on a given TiB of storage is halved with each doubling of Netspace[0][1]. The current return is ~$1.50/TiB/Day. As the expected return declines, farming CHIA will become less attractive than other crypto 'investment' opportunities, and the whales will turn their attention elsewhere.

I'm guessing that netspace starts to level out after another doubling or two--assuming the exchange rate remains the same.

[0]https://www.chiaexplorer.com/charts/netspace [1]https://www.chiaexplorer.com/charts/xchTib

[22c]

>The current return is ~$1.50/TiB/Day.

I'm bearish on Chia, but I can see why it's attractive at this price point. I thought the profitability was much worse, but I did the calculations myself and it looks like around ~1.30/TiB/Day.

At the $20/TiB price point for HDDs that seems like you'd be in the money after ~16 days (excluding plotting costs and farming power usage). The big problem is the lottery, which could be smoothed out by pool farming.

Yes, the rewards are getting worse by the day, so it's hard to predict where things will be at in 16 days time, but again, pool farming should theoretically smooth that out a lot.

[imtringued]

Yeah, the problem is actually worse. Bitcoin is energy intensive. Chia is capital intensive. Producing capital costs energy so it effectively boils down to proof of work (burning energy to produce HDDs and SSDs). However, scaling energy is very easy, scaling HDD production is very difficult.

The only benefit is that storage can be repurposed very easily, thus booms and busts may cause temporary oversupply of storage.

[dheera]

Chia isn't worth farming anymore.

I have a pile of old spare 512GB SSDs and it looks like I would able to make something like $2/day. Not even enough for a coffee.

Same with Ethereum. Tried to mine with my Titan V. 31 MH/s. Also not even enough for a coffee.

This is meaningless, and cryptocurrencies are effectively centralized around mining farms with ASICs.

[imtringued]

I bought a 512GB SSD for around 60€ some time ago. Let's be generous and assume 40 days pay back. It could be 90 days and it still would be worth it. I don't know how fast SSD wear out when used for Chia though, it could be unprofitable for all I know.

[dheera]

I'm 90% sure as of this week you can get more money by selling the SSD than farming Chia on it before Chia wears it out.

[ipsum2]

I don't believe that an ASIC has been developed for Ethereum, but I'm not too familiar with cryptocurrency.

[jl2718]

Long ago. See the hash rate chart and guess when that might have happened. Hint: long before they were offered for sale.

[devwastaken]

That's not how chia rewards you though. It's a lottery, and though chia may be around 1000 or so right now it could massively increase in the future. Current projections are something like 3-4 months to get a single reward. But if you could make potentially an extra 50K a year with some harddrives that's not a bad deal.

[hellow0rldz]

BitCoin mining is also a lottery. I could stumble on a good hash using my own CPU. But the odds are in favour of the farms with ASICs.

[hi5eyes]

btc can only be mined on asics

[22c]

Simply not true.

https://gizmodo.com/mining-bitcoin-with-pencil-and-paper-164...

[dheera]

That "3-4 months" is going up at an exponential rate, and it's so bad now that by the time 3 months comes by the expected time to reward is 6 months. Even if you get 1 reward, you won't get a second reward unless you continuously keep buying drives and racks for them. You can't really get continuous passive income on a fixed set of hardware.

50K/year is impossible at this point with a consumer rig.

[lozaning]

> Current projections are something like 3-4 months to get a single reward

You cant project time to win without knowing how many plots a farmer has. You've gotta have 100+ TB plotted (prolly closer to 140actually) at this point to have a chance at winning every 3-4 months.

A single plot would put you closer to 7 years.

[kelp]

I currently have two plots farming and it looks like this:

  Farming status: Farming
  Total chia farmed: 0.0
  User transaction fees: 0.0
  Block rewards: 0.0
  Last height farmed: 0
  Plot count: 2
  Total size of plots: 202.674 GiB
  Estimated network space: 7357.090 PiB
  Expected time to win: 20 years and 1 month

[tomlagier]

That's just not accurate. See https://chiacalculator.com/ for more.

3-4 months is currently around 18 TB.

EDIT: 100 TB would get a coin every ~ 18 days on average.

[dheera]

> 3-4 months is currently around 18 TB.

That number does't account for how many more Chia farmers will exist in 1,2,3 months. It doesn't even account for linear extrapolation of total plots in the world, let alone exponential.

[imhoguy]

Not good indeed. Small hosting businesses and consumer storage will suffer, because big cloud companies usualy have HW production customized and contracted ahead.

[ulzeraj]

And you didn't even touched on the e-waste issue.

[inshadows]

Right. Your cloud rents will rise by few cents.

Seriously, anyone compared the capacity of Chia and beloved cloud providers?