Bitcoin’s achilles heel — or why it can’t be the new gold

So the current zeitgeist is that bitcoin is the new gold.

The proponents say the divisibility, fungibility, scarcity, and decentralized properties are similar to gold with the added advantage of it being digital. While all that seems true, I think those are only the surface level similarities.

Why bitcoin is different

The main difference is right in the original bitcoin whitepaper: Bitcoin’s “mining” and “transacting” functions and inexorably linked. Unlike gold, where a “miner” has no power over who can trade gold and how much it costs, the bitcoin miner can, which leads to some interesting conclusions.

If I have 50% control of all the gold mines in the world, I can’t actually affect any of the trading that happens in the world. However, if I own 50% of the bitcoin hashing power, then I can choose which transactions to allow, and even cause “rewrite” older blocks in the blockchain (it gets more difficult the more blocks I want to rewrite)

You can’t move any money between wallets, but someone with 50% hashing power can potentially cause irreversible havoc on the blockchain, maybe even rendering bitcoin worthless (Is it worth anything if you can’t transact it?)

I think this link between the transactions and the mining(aka the blockchain) is the Achilles heel of bitcoin. This is the fundamental difference between gold and bitcoin that may not allow bitcoin to replace it in a meaningful way.

How it can happen

While initially buying bitcoin, I did the following thought experiment in mind of how much bitcoin could be worth:

If bitcoin has an active shorting market, and following that the hypothesis above holds ( you can make it drop in value if you control >50% hashing power), then all you have to do is short bitcoin and purchase 50% hashing power. As long as you can make your bitcoins value drop more than you paid for the compute, then you’re in the green.

Some back of the envelope math:

- Current bitcoin hashrate (https://www.blockchain.com/charts/hash-rate ): ~140 million TH/s
- Cost for hardware: (https://en.bitcoin.it/wiki/Mining_hardware_comparison) — This is a list of commercially available miners, I have no doubt someone with access to a fab or buying in bulk would get a significant discount): The best price/performance on that is 6.4 GH/s/$.

To get 70m TH/s, you would need to spend approx 70 * 10⁶ / 0.0064 which is approximately ~11B USD. There’s the cost of electricity as well which I haven’t added here but that would be dependent on the amount of time you wanted to carry out the attack (Quick research says it could be as low as $4M USD / day https://www.frontiersin.org/articles/10.3389/fbloc.2020.565497/full)

But 70m TH/s would only give you 70/(140 + 70) = 33% of the mining power. To get to 50%, you’d need the full 140m TH/s so that would cost approximately 22B USD. Maybe you can get by with only 33% of the hash rate (https://arxiv.org/abs/1311.0243)

Right now, I don’t think it’s reasonably possible to short >$22B USD of bitcoin, but if bitcoin’s price keeps rising, there will probably be a time where someone with deep pockets can carry out such a trade — in which case it feels like bitcoin is vulnerable.

If we had perfectly liquid and efficient markets, bitcoin can’t be worth more than this this amount (~22B USD), because there’s free money to be made under this hypothetical scenario: short all the bitcoin, launch a 51% attack until the value drops to 0, then buy it all back and you made back the money spent on the hardware.

This is obviously a naive view under a perfect world, but I don’t think the numbers are more than an order of magnitude off from reality. Even if bitcoin’s value drops only by half, there’s still a profitable trade to be made if you buy enough. I’m not even considering other actors like governments who may want to do this even if it’s unprofitable.

As of today (3rd Jan 2021), bitcoin’s market cap is $628B USD (https://coinmarketcap.com/) so you’d only need to short ~4% of it. The more you could short, the better; and the less the value has to drop to make an attack profitable. You’d have to spend even less if you can get by with a 33% attack or purchase hardware cheaper.

In comparison, at its peak Tesla had 18% of its shares shorted (https://markets.businessinsider.com/news/stocks/tesla-stock-most-shorted-companies-us-traders-betting-against-apple-2020-2-1028873641). With more mature financial markets around bitcoin, shorting 10 or even 20% is not unrealistic.

Under this theory, bitcoin’s value can never be more than the hardware required to mine 50% of it i.e — the price of hardware today for its current hashrate.

Also assuming that $/TH/s keeps going down over time (Moore’s law and all that), there has to be a constant increase in $ spent in hardware just to maintain the price of bitcoin. In other words: if bitcoin price is stable, and hardware keeps getting cheaper, eventually this is a no-brainer for someone to do this.

The situation only worsens when there is no more bitcoin left to mine. The transaction fees alone have to be enough to incentivize the miners not to coordinate a 51% attack as they can’t sell any mined BTC (but they could easily short 5% if bitcoin)

I also haven’t tracked the historical market cap : hardware cost ratio, but it would be interesting to see. Eventually it may be big enough for this to be an obvious trade. The only question is whether that is 100 : 1 or 1000 : 1

Objections / Questions

Bitcoin’s value will not go to 0 with a 51% attack:

— Maybe the bitcoin holders can do a counter-attack by buying even more hardware. But the equilibrium here is still where there’s enough of a hashrate so that the attacker/short will not have a profitable trade which returns us to a hardware cost = bitcoin value.

— The attacker can’t and will not hold on to the attack forever so all you have to do is wait it out: That’s true, but once the genie is out of the bottle, others may try it. The attacker can just as easily just start mining bitcoin with the hardware — so maybe it isn’t even a sunk cost. It’s easy to use the hardware to just go between attacking / mining at any time.

— The price doesn’t have to drop to 0, it’s just a function of how much you can short. If you can short ~44B of BTC in the scenario above, then it only has to drop 50% in value for a profitable trade.

— People can still transact by sharing their private keys to their wallets off the blockchain by some other means, so you can’t completely stop bitcoin transactions from happening: That’s also true, but there are other nefarious things an attacker can do such as rewrite the blockchain more than enough blocks to void “confirmed” transactions and so on.

— 51% attacks have never happened in the past: They do seem surprisingly rate. Maybe the coordination is too hard, or bitcoin has escaped the threshold where any one person / entity has the cash to do this individually.

Crypocurrency in general can find a way around a 51% attack.

— I haven’t read enough about this other than proof-of-stake coins which may avoid this better. Since the holders of the coin have the power here, there isn’t really way to attack it other than purchasing 51% of the currency yourself — in which case the attack above fails. Maybe there are other ways around the 51% attack in a Proof-of-Work cryptocurrency.