Nostr Web Client

talk is cheap man, lets see some numbers

Aggregate CPU Power: Datacenters vs. End-User Devices

When considering the aggregate CPU power of datacenters versus end-user devices in the U.S., the disparity is striking. Here’s a breakdown of the aggregated computational power of both setups.

Aggregate Power of Datacenters

Total CPU Count:

A typical large datacenter can house thousands to tens of thousands of servers. Assuming an average of 16 CPU cores per server, this can translate to hundreds of thousands to millions of CPU cores.

Performance Estimation:

If we estimate a conservative average of 10 teraflops per server, a datacenter with 10,000 servers could achieve:

10,000 servers × 10 teraflops/server = 100,000 teraflops (or 100 petaflops).

Higher Scale Examples:

Major cloud providers like Amazon Web Services (AWS) or Google Cloud can aggregate even greater power, reaching 1 exaflop (1,000 petaflops) or more through optimized hardware and workload distribution.

Aggregate Power of End-User Devices

Total CPU Count:

There are approximately 300 million personal computers (desktops and laptops) in the U.S. Assuming an average of 4 CPU cores per device, this results in about 1.2 billion CPU cores.

Performance Estimation:

If we take an average performance of around 500 gigaflops per device, the aggregate computational power would be:

300 million devices × 500 gigaflops/device = 150,000 teraflops (or 150 petaflops).

Summary of Aggregate CPU Power

Datacenters: 100 to 1,000+ petaflops (potentially reaching exaflops)

End-User Devices: 150 petaflops

This aggregate comparison clearly illustrates that datacenters collectively provide an order of magnitude more computational power than all end-user devices combined. The architecture and capabilities of datacenters offer substantial advantages for large-scale processing needs, whereas end-user devices serve general consumer purposes.

Chatgpt

Reply to this note

Please Login to reply.

Discussion

Hanshan 5d ago

well I guess I asked

but minus 10 points for the low effort AI slop

so taking the hive mind at face value

we're left with the question is it easier for an adversary to either a pay or force 10-50% of Amazon and Google data centers to cooperate

or to coerce Bitcoin miners.

satoshi jr 4d ago 💬 3

Magnitude is all you need to know.

You have 2 problems.

1 if the government can win the mining with current hardware and energy by an order of magnitude in the best case scenario. And the concentration is probably increasing. You can't have decentralization.

2 you are in the timeline where bitcoin exists and is several orders of magnitude bigger. You don't even have the option of coopting government power like the way mining is getting integrated into power grid stsblization.

Hanshan 4d ago

you don't have decentralization of Bitcoin miners either, as they are ALSO concentrated in a handful of large registered businesses. its the same picture.

except having *potential* hash divided up among large entities who profit from it being used for different, varied functions is obviously MORE decentralized, not less.

and its just fantasy that Bitcoin miners are "co-opting government power" to any significant degree by providing grid stabilization.

satoshi jr 4d ago 💬 2

Hash Power in Large Public Companies vs. Total Hash Power

As of the end of 2025, public companies engaged in Bitcoin mining represent a significant portion of the overall computational power available for mining, but exact comparisons vary.

Total Hash Power Overview

The total global hash power is estimated to be around 350 EH/s (exahashes per second). This value represents the combined computational capability dedicated to Bitcoin mining fundamentally across various industries, regions, and private operations.

Contribution from Large Public Companies

Large public companies, such as Riot Platforms Inc., Hut 8, and Marathon Digital Holdings, contribute a considerable share to the total hash power. Reported figures suggest these companies, in combination, account for approximately 20% to 30% of the total hash power. For example, Riot Platforms reportedly had around 8 EH/s dedicated to Bitcoin mining, equating to a significant fraction of the overall total.

Here's a breakdown of some key players and their contributions:

Riot Platforms: ~8 EH/s

Hut 8: ~4 EH/s

Marathon Digital: ~7 EH/s

Implications and Trends

The involvement of public companies in Bitcoin mining has implications for both scalability and resilience in the network. Their larger operational scales allow for capital investment in cheaper energy sources and advanced technologies, essential for maintaining competitiveness.

Moreover, there's a notable trend towards diversification, with companies pivoting more towards data center operations, blurring the lines between traditional mining businesses and tech-based solutions. This transition is critical as profitability from mining becomes more volatile due to fluctuating Bitcoin prices and rising operational costs.

Conclusion

In summary, while large public companies make up about 20%-30% of total hash power, the ongoing changes in the sector may lead to shifts in this dynamic. The rise of energy-efficient technologies and enhanced operational strategies will likely continue to influence the landscape significantly.

Very different landscape and magnitude in terms of concentration risk

Hanshan 4d ago

one

I'm entertaining your AI slop for the sake of conversation.

but I don't really believe it knows wtf it's talking about or accept it as a convincing argument.

two

you repeatedly gloss over the fact Bitcoin miners are known and visible. they are public and operating.

speaking about *potential hash* in data centers is not the same thing. that hash has to be acquired and brought online. these are not the same thing and it isn't particularly useful to compare them as if they were.

you want to talk as if bitcoins active, known, centralized hash can't be effected by well understood and straightforward regulatory pressure.

but somehow it's trivial for an adversary to bring online against monero *somebody else's compute* that is already being used for profit in another way.

so even if we accept that the AIs numbers are correct, and I don't, they aren't directly comparable at all, as they aren't a measure of anything that can be called "decentralization"

satoshi jr 3d ago 💬 2

One you dont like the slop do your own research

Two yes you are right the large public miners are known and will be used by the government to attack bitcoin.

They will control sometgin in the magnitude of 20-30% maybe its as high as 50or 60? dunno

The problem is that the same attack on monero means that there is 10x more hashnpower ready to go against all ofnthe potential hash power of all desktop and laptops if they all only mined. Maybedl its only 2x the amount.

But it's not 2x current monero hashing.

Its 2x all potential mining

Its Already lost

Hanshan 3d ago

except that hash is not "ready to go." it's already being used for other profitable purposes. you ignored the point of the post, which is about the friction of bringing pressure to bear at all.

You're trying to pretend like the US government is a single cohesive entity that can just call Amazon and use all of their compute power for whatever it wants. I'm pretty sure it doesn't work like that.

but I don't know and neither do you. which is how I started the conversation.

and you're just making up numbers.

while I agree that there is certainly vastly more potential hash out there that could be mining monero, you continually ignore the fact that they don't need to physically control Bitcoin miners at all. they just have to send letters.

nothing has already lost. another topic we haven't touched on yet is *the desire to attack the chain in the first place*

this is another area where Monero has an advantage because of bitcoins visibility.

so to come back to the beginning

its trade-offs. both approaches have their negative points.

satoshi jr 3d ago 💬 2

When bitcoin or monero becomes an existential risk to the us government they will taken over any industry and try to destroy you.

In that situation you should fully expect all public miners to be seized and any potential miners to be used against you

If you're hoping laws will protect you, you're an idiot

Hanshan 3d ago

you'll talk about anything EXCEPT the friction involved in trying to hijack privately owned compute wont you?

obviously the threat to Bitcoin is higher.

because Bitcoin has a higher fiat value and is more visible.

because it's easier to put pressure on miners.

I agree that if Monero ever becomes an existential risk they'll do whatever it takes. I hope that becoming an "existential risk" means price appreciation and therefore greater mining profitability.

and at that point you have higher floor of decentralized compute power protecting the network

and the centralized compute power is more incentivized to protect the network rather than attack it. just like with Bitcoin.

but the game theory of that particular situation depends on specifics that we can't know.

Moral of the story is that *right now Monero is in a much more secure place than Bitcoin is vis a vis mining centralization.

satoshi jr 3d ago 💬 2

Its very poor risk analysis if all you're looking at is what's happening right now. Its like saying driving without a sear belt is fine because i haven't died or been in a car accident yet.

You try to look for worst case scenarios and how the structure of things create risks.

I agree that monero is too insignificant to be attacked right now, but after bitcoin already got big people already know that monero could get big enough to be a threat.

And there's orders of magnitude of what they need to attack monero just sitting there already.

Hanshan 3d ago

for the last time, IT ISN'T "JUST SITTING THERE."

they have to do something in order to acquire it and use it to their advantage. that is friction that neither of us can accurately evaluate.

and I didn't say you should ONLY look at how things are right now.

I pointed out thats the current situation and how things are in the future depends on specifics that neither of us can be confident about.

satoshi jr 3d ago 💬 2

OK you're really not getting this. There's effectively 0 friction because when you get to the point where when they want to take it you can not physically stop them.

That is the physics of the situation

You're right that before that point you can't know exactly how the theater of lawfare will be done.

It doesn't matter.

The way in a car crash it doesn't matter if your licence is up to date or if the other guy did an illegal turn. None of that changes physics

There's enough compute right now to fuck up your network. Its order of magnitude more.

When they attack bitcoin they physically cannot get an order of magnitude more of hash power.

Do you see how one is a physical limitation and the other is just legal and economic.

Those are not the same category of risk

Hanshan 3d ago

that does help me understand what your point is, thanks.

but I still disagree that it means that there is zero friction.

it's true that legal protections are not to be counted on. but neither are they completely irrelevant. whoever the adversary might be (and without knowing who that is we can't calculate risk can we?) isn't just co-opting Google and Amazon compute arbitrarily.

I understand that you mean that there is *potentially zero friction*

at the point where they have taken the gloves off and want to destroy you.

and that's worthwhile knowing. it's just not the actual situation right now.

but it's certainly worth reflecting on in the Monero community that, if compute is as centralized in large data centers as you say,

then regardless of how many plebs are mining themselves, the network is still vulnerable.

satoshi jr 3d ago 💬 3

Again the point of the risk analysis is not to look at when things are going well. Instead to consider the potentially catastrophic situations and your exposure.

You dont calculate risk that's how you get 2008.

You cant know what will happen but you can try to understand the structure of how things are.

With enough time you will hit all of the extreme situations and the only this that matters is if you can survive

Hanshan 3d ago

nobody said that.

risk analysis is not just looking at the worst case scenarios either. it's about making intelligent decisions based on the available information.

I don't live in a bunker in Idaho because there *might be nuclear war.

and sometimes existing in the shadows is a better long-term survival strategy then getting Wall Street involved.

satoshi jr 3d ago

No risk analysis is about analysing risk. It dosnst mean you can't choose to accept the risk.

You are choosing to take nuclear risk

The us government doesn't think that's acceptable so they have bunkers.

The point of this conversation was about money. You can't do a shit job of hiding in the shadows and claim you're gonna be better money.

satoshi jr 3d ago

Oh and you are implicitly saying that you're only looking at rosy scenarios if you aren't actively looking for the extreme situations that are exisistntial risks.

But again its only if you're trying to be money not just a tech demo

satoshi jr 4d ago

Also your stipulation that having alternate uses for your compute decentralizes is probably not true.

The alternative uses create the large concentrations in the market already. Because data centers are useful closest to end users they are located closer to major urban centers, is centralized for efficiency.

Ercot already has miners as part of their load balancing. I think it was 10% marginal load but I dont remember.