An Interview with Andreas M. Antonopoulos regarding 51% attacks on Bitcoin

[Pmalek]

In one forum post the other day, I stumbled upon a YouTube link that a user posted in which Andreas M. Antonopoulos talks about 51% attacks on the Bitcoin network. It is a scenario that seems possible to many, but Andreas perfectly explained why such an event would be difficult to carry out.

When he visited Berlin for a crypto conference in 2019, he gave an interesting interview which you can watch here:
https://www.youtube.com/watch?v=7w-Q2GOZJSQ

Here are the most important parts that were mentioned in the interview:

Andreas was asked whether it would be possible for a nation state to acquire a sufficient amount of hardware to carry out a 51% attack on the Bitcoin network, and to completely take control of the whole network.
Andreas considers that an impossible task, and gives several reasons to why that is.

- The required amount of mining hardware needed for a 51% attack on Bitcoin is simply not publicly available for sale. Companies that manufacture and sell such mining devices, are already doing so at full capacity. Everything they manufacture, they sell very quickly. There are no hidden stocks that someone could buy up. He mentions an interesting fact that if someone wanted to carry out such an attack, he would have to constantly buy all the ASIC miners from all of the companies that exist on the market, for more than a year. With the sole aim of getting close to that 51% hashrate needed for an attack. So this is an impossible task.

- The second assumption is that someone could buy the required materials, needed to produce a sufficient amount of mining devices for a successful attack. Andreas says that the best miners are made of 16-12nm silicon chips. Only those made with the best quality materials would have to be used to reach that rate needed for a 51% attack. If chips with lower quality materials were to be used, many more mining devices would be needed. The companies that produce this type of silicon, produce it mostly for graphics cards and devices for artificial intelligence units. They already have plenty of work to do and don’t have the capacity to meet the demands of a new client who would like to buy all of that to attack the Bitcoin network.

- The third possibility that Andreas talks about is that someone builds their own factory where they would produce the required devices. The investment needed for such a venture would be tens of billions of US$. These devices would then consume a huge amount of electricity and create new expenses in the multi-million dollar regions.

Even if such an attack were to be successful, it must be sustained. And that requires additional huge investments. Now, here is the problem. Those who mine Bitcoin honestly, get rewarded for their work with mining fees and newly created bitcoin. They cover their expenses (for electricity) and make a decent profit. Those who attack the network don't receive any compensation from the network because they don't include transactions in their newly created blocks. They must cover their own costs and maintain the attack. Andreas estimates that these costs amount to 2000-3000 BTC per day!

Bitcoin created in this way would ultimately become worthless. Exchanges could blacklist it and it would never be sold. The attack could be resolved with a fork of the network, where the miners who respect the rules just continue with their work, while the other side that does not respect the rules and creates empty blocks would simply be excluded and blacklisted.


The end result of all this is tens of billions of US$ spent on the purchase and manufacturing of mining devices, and millions of US$ of electricity costs for an attack that could be quickly discovered and resolved with a fork. The miners who attack the network can only prevent the confirmation of certain transactions, or carry out double-spend attacks. In both scenarios, those coins could be blacklisted and would have no value. The Bitcoin network would not be interrupted. It would become even stronger over time because it overcame and survived a 51% attack.

In the second part of the video, Andreas talks about the differences between miners and full nodes.


Have a look:
https://www.youtube.com/watch?v=7w-Q2GOZJSQ

[Husires]

51% attack of the network is guarantees the hacker will do double spending with a high probability of sucess given that he has the longest chain, but someone with less than that percentage may try to try luck, but he will need to be lucky to succeed in the matter so the current hash rate of the network makes it protected Of such attacks, which may happen in many altcoins.

Anxiety is not from such attacks, but rather from scalability

[hatshepsut93]

Bitcoin created in this way would ultimately become worthless. Exchanges could blacklist it and it would never be sold.

Andreas is conveniently forgetting that mixing is a thing, and that to this day criminals in many cases manage to effectively mix and withdraw their dirty coins.

The attack could be resolved with a fork of the network, where the miners who respect the rules just continue with their work, while the other side that does not respect the rules and creates empty blocks would simply be excluded and blacklisted.

How do you distinguish between "bad" and "good" miner from a full node's point of view? Blacklisting some IP addresses obviously wouldn't work.

The only thing that prevents 51% attack is the sheer amount of hashing power, if that would be to change in the future for some reason, then 51% would be more likely. Look at altcoins, it already happens to them, and no mechanism aside from hash power is able to stop it.

[erikoy]

I am not a techie when it comes to blockchain but as I read through the OP, it made me some relief that bitcoin will stay strong after such attack and that hackers should be able to sustain their activities while spending lot of money. I do believe in bitcoin system but with these attempt of hacking it may also create another threat to bitcoin system. They had already started it and may continue to explore other ways to hack the system. They may not be successful now but who knows in their next attack it would be successful and probably bitcoin will fall which we do not like it to happen.

[Sled]

Andreas says logically what he believes, unlike others. Maybe he is right or maybe wrong but what I take to this is that 51% attack on Bitcoin is just like politics. Certainly not a doing of a single person but consist of many individuals as a group. I'm not saying that could be possible to succeed coz I believe it won't, and Bitcoin will remain decentralized no matter what it takes.

[Csmiami]

How do you distinguish between "bad" and "good" miner from a full node's point of view? Blacklisting some IP addresses obviously wouldn't work.

Taking a wild guess here, so please anyone right me if I'm slipping; but pushing a Core update (say 0.20.2) where the "bad" miners are blocked or ignored, and having nodes migrate to that wouldn't do the job?

[hatshepsut93]

How do you distinguish between "bad" and "good" miner from a full node's point of view? Blacklisting some IP addresses obviously wouldn't work.

Taking a wild guess here, so please anyone right me if I'm slipping; but pushing a Core update (say 0.20.2) where the "bad" miners are blocked or ignored, and having nodes migrate to that wouldn't do the job?

How do you even identify the miners? Right now we only know who mined which block because they want us to know. Miners can easily use ToR, mixing and other privacy tools to make it impossible to identify them, even if they are behind 51% of the hashrate. By design Bitcoin's protocol knows very little about the identities of its participants, and you can't selectively change it without ruining whatever privacy we have for everyone.

[DooMAD]

I'd still prefer if we could move away from the name "51% attack".  It's a misnomer that leads to new users going away with the wrong impression.  Hashrate attacks can be performed with less than 51% of the hashpower.  So just call them "hashrate attacks".

[stompix]

- The second assumption is that someone could buy the required materials, needed to produce a sufficient amount of mining devices for a successful attack. Andreas says that the best miners are made of 16-12nm silicon chips. Only those made with the best quality materials would have to be used to reach that rate needed for a 51% attack.

That was years ago, S15 launched in 2018 had BM1391 7nm chips, both Bitmain and Cannan are testing 5nm designs by now, and at least Bitmain is claiming to have a functional prototype, a bit weird hearing such a mistake from him.

The end result of all this is tens of billions of US$ spent on the purchase and manufacturing of mining devices

Nope, not even close to billions.
The current hashrate is around 130 Exa, you need 1.1 million S19pro for this, bitmain was telling them for 2500$ with a profit so you could get it for about two billion, that's 1/6 of the annual defense budget of Sweden....or 0.1% of the F35 program    Not even mentioning the fact that such a sudden increase in hashrate would put easily legit miners out of business and it will give you also a bit of reward while mining to cover costs.
Hard to do it, yeah, maybe. Impossible? Let's serious, if a medium GDP-sized country would have really wanted that it could have done it and it can do it.

Those who attack the network don't receive any compensation from the network because they don't include transactions in their newly created blocks. They must cover their own costs and maintain the attack. Andreas estimates that these costs amount to 2000-3000 BTC per day!

That's stupid, they receive the block reward, fees are barely 10% of the rewards, if the cost of 51% attack would require 3000 BTC what were the miners mining for when the entire daily reward was before the halving at around 1800+- 10%?

Seriously I'm disappointed hearing this from him, he is focusing on double-spending rather than crippling the network.

Take a look at the mempool, we're averaging 10% slower blocks right now, that would be like a miner mining 10% of the blocks but all empty, and look at what the fees are doing, lowest fee in 5 blocks was 87sat/b, now imagine 30%!

In both scenarios, those coins could be blacklisted and would have no value.

The moment you start blacklisting en-masse something on the blockchain you can stop calling it bitcoin.

[zasad@]

Thank you for the good material, added the translation to the Russian board
https://bitcointalk.org/index.php?topic=5211294.msg56357130#msg56357130

[Pmalek]

Bump