Add an example calculation for cost-to-attack

[KevinMusgrave]

https://twitter.com/LightningSource/status/1560304069494951945

[KevinMusgrave]

Text:

Assume a competitive market where profit margins are 0%.

Assume ASICs last for 3 years. So every year, 1/3 of ASICs have to be replaced.

Assume miners want infrastructure (land + buildings) to be paid off in 20 years.

At 0% profit, Revenue == Cost Annual revenue = $17 billion Annual cost = (ASICs/3) + (Infrastructure/20) + Energy

Assume Infrastructure = $20 billion Assume ASICs = $30 billion

$17 billion = ($30 billion/3) + ($20 billion/20) + Energy

Energy = $6 billion

An attacker needs:

• 50% of ASICs, costing them $30 billion. (The total cost of all ASICs becomes $60 billion, which they now have half of.)
• 50% of energy, costing them $6 billion.
• 50% of infrasturcture, costing them $20 billion. So the total cost is 30 + 6 + 20 = $56 billion

This is definitely something a government, or multiple governments working together, could afford. Actually at these costs, they could acquire much more than just 50% of hash power.

They wouldn't be doing it for the sake of double-spending. Most likely they will want to either control bitcoin, or cause it to fail by decreasing confidence in its security. And at a $20 trillion market cap, it becomes a very attractive target.

[KevinMusgrave]

Other points:

• It doesn't have to be the USA that attacks bitcoin. It could be China.
• Resources can be acquired over several years:

  
  Maybe they buy $5 billion / year. So it'll take them 6 years to get to $30 billion:

Year 1: 5 Year 2: 5 Year 3: 5 Year 4: 5 + 5 to replace year 1 Year 5: 5 + 5 to replace year 2 Year 6: 5 + 5 to replace year 3

Total ASIC cost = $45 billion instead of $30 billion

- Why assume 0% margin? If I assume a profit margin greater than 0%, then the cost to attack decreases. For example, if I assumed a 50% margin, that means total costs are 17/2 = $8.5 billion. That would be easier to attack. So I assumed 0% because that is the best-case scenario in terms of security.