Hacker News Reader

Comment by spir | Hacker News Reader

Note that Ethereum is switching to proof of stake this year, almost surely before Q4. It's true that this upgrade has been coming for years, and it's also true that it's actually arriving this year (if interested, see https://eth2.news).

After Ethereum switches to proof of stake, global demand for GPUs will plummet as GPUs are uncompetitive for BTC mining, and beyond BTC, ETH is by far the largest remaining proof of work coin.

View on HN

Note that Ethereum is switching to proof of stake this year, almost surely before Q4.

Is there some way you can bet against that claim?

Absolutely! Please see this Polymarket prediction market on "When will Ethereum switch to proof of stake?"

https://polymarket.com/market-group/ethereum-merge-pos

Rest assured there are many gamers who will pick up the slack, not to mention businesses who want them for training models.

This. I've been planning to buy a new gaming rig for two years, but wasn't so much a priority that I would pay exorbitant prices for even midrange hardware.

When normality is restored, I'll be happy to finally upgrade.

I will be one of them :)

But people who don't necessarily need it and/or don't make money with it won't pay insane prices for (ab-)used mining GPUs. It won't be insanely cheap, but definitely a lot cheaper.

When the time comes, for a casual gamer which popular etehrum mining GPU’s would you be on the lookout for?

The main difference is that gamers are consumers, while ETH miners use GPUs to make income. Even though gamers were buying scalped GPUs, I'd imagine GPUs going to back to being treated as a consumable (rather than something that generates income) would help prices.

Companies using GPUs to train models are also using them to make money. They will pick up the slack the miners leave behind.

I think you're overestimating hoe many companies use GPUs to train models, I highly doubt there are enough to drive prices up the way crypto had.

Aren't Quadros/FirePros with lots of VRAM better for that?

I wonder what will happen: Ethereum switching to PoS or Tesla having full self-driving.

People that care about PoS could have already invested in something else. The people stuck on Ethereum will boast about PoS coming soon, but if tomorrow they said they were going to wait for 5 more years they'd stick with Ethereum and talk about PoS for 5 more years without switching to a different PoS token.

There are a handful of other PoW coins not far behind Ethereum so I wouldn't count on demand plummeting

https://whattomine.com/

Can you translate for people (like me) who don't know what all those numbers are supposed to represent? I can't make sense of it or relate any of them to GPU demand.

I don't understand much of the site either, but at a high level, my guess is that it's sorting crypto coins by computation required vs $$$. Right now ETH is at the top, aka for every flop you put it, you get the most return, but there are many others not far behind. So even if you were to eliminate Etherium entirely, there would still be plenty other that are profitable to mine.

Why sell a GPU for a fraction of the price, when you can just switch over to another coin and make 50-90% of the money you were making still.

If you have a GPU you could use it to "mine" various coins. That site calculates which of them is currently the most profitable and how much profit you're likely to make with a given hardware configuration

Thanks, I hadn't seen that site.

In my view, this site is strong evidence for the thesis that demand for GPUs will plummet, here's why:

1) If we tweak your link to sort by Market Cap, we can see the next biggest coin to mine after ETH is has only 1% of ETH's market cap https://shorturl.at/lxLS5

2) These smaller coins do not have the capability to automatically or naturally grow bigger market caps, trading volumes, and buy-side demand just because ETH switches to PoS. Mining profitability ultimately comes from a coin's buy-side demand, which is proportional to many things-- fame, developer activity, etc.-- and none of the PoW chains in that list are anywhere remotely close to "breaking out" and growing proportionally with Ethereum, or, say, Solana. How do I know this? Well, I'm an insider that focuses all of his time on ethereum and web3, you'll have to take my word for it :)

As someone rather naive in regards to cryptocurrency, I don't understand what you're basing this on:

> These smaller coins do not have the capability to automatically or naturally grow bigger market caps, trading volumes, and buy-side demand just because ETH switches to PoS.

What would they need to cross that barrier? For example, don't all coins have the ability to manipulate trade volume by sending money between owned wallets?

Sending money between two wallets you own creates transaction volume not trade volume. The latter comes from orders matching against each other on an exchange and requires paying the exchange fees so it is not as easy to fake (you need the exchange itself to be in on it, or they take all your money in fees).

The combination of trust me I'm an expert while not knowing about #1 site for coin mining profitability...not a good look

Yeah, its not gonna plummet. Theres still going to be huge demand. PS5s and Xboxes arent used for crypto and theyre still sold out

Right but what happens when another blockchain technology hits the scene that’s trendy and happens to be PoW? People who made significant investments into GPU miners and the infrastructure to support them won’t go away quietly.

That's fair, what if another PoW coin gets really popular?

The good news here may be two-fold

1) the existing group of public chains have matured greatly and their network effects are quite strong now. It is no longer so easy to start a new chain

2) virtually all new chains are PoS because it's understood to be significantly more secure and less expensive vs. PoW

3) public chains take years to grow, and right now, there is no PoW public chain on my radar that has a chance of generating nearly as much mining demand as Ethereum (and Bitcoin but ASICs)

For those reasons, the idea of a new public chain appearing from nowhere and impacting GPU demand seems in practice unlikely to occur.

> virtually all new chains are PoS because it's understood to be significantly more secure and less expensive vs. PoW

PoS security relies on the integrity of the stake — and the integrity of the stake relies on the “security” of the PoS system. It’s an intractible circularity problem.

Yes, a large sample of recent altcoins have shipped PoS, but this is no more relevant to a PoW v PoS security debate than a modern top 40 “hit songs” chart is to a debate over Beethoven v Bach.

>and it's also true that it's actually arriving this year

Do you have insider knowledge or just a quality crystal ball?

I have been full-time on Ethereum for years, I'm a software engineer, and I have a relatively intimate knowledge of the specific milestones the community has been achieving and why that implies that proof of stake will ship this year.

Here's an overview

Dec 2020: proof of stake launched in production, but isn't yet used to secure ethereum, it runs in parallel

October 2021: the last major upgrade to Ethereum before switching to proof of stake went live in production

Today: proof of stake "Merge" testnets are live, all core teams are 100% focused on the merge, regular progress reports are nominal, etc.

What's the difference this time?

Please see my other answer to a sibling comment https://news.ycombinator.com/item?id=30490897

Any insight why it was postponed?

Reply on news.ycombinator.com