I worry that a 51% attack on Bitcoin is inevitable. It is frightening to think that an organization could buy up enough mining equipment to fork the blockchain.
Right now let's say that KNCMiner had ample supply of their KNC Neptunes -- one of the best rigs as far as cost/hash. This is a hypothetical because KNC has sold out on what they accepted so far for pre-orders for future shipments. But let's say you could place a large order and get it shipped overnight (or at least, before difficulty jumps again where you need to order more).
Total network capacity (currently, using BitcoinWatch.com): 16,116 Th/s (SHA-256).
KNC Neptune performance: 3 Th/s.
KNC Neptune cost; $10K
So you would need 5,372 of those bad boys to get 51%. That's a $54 million investment in hardware.
Now let's say you had that capacity online today -- mining secretly on your private fork of the blockchain.
What capabilities do you have?
You can send bitcoins to an exchange(s), and then on your private fork of the blockchain you can spend those same bitcoins elsewhere (to another address you control, for instance). But to gain any value you need to take delivery of something of value in a non-reversible manner so that after you broadcast blocks from your private fork the exchange(s) you cheated have no recourse.
How are you going to get $54 million worth of value from an exchange(s)? Withdraw Bitcoins? Most exchanges emply hot wallets with a limited balance (with the remaining coins held in cold storage) -- there's no way tens of millions of dollars worth of non-reversible funds will be available for withdrawal immediately. Even if there was, they wouldn't be available for withdrawal anonymously.
So there's really no economic incentive for performing a 51% attack for the purpose of double-spending against Bitcoin.
That doesn't mean it won't or can't happen, it just means the likelihood of it occurring is pretty small.
Now take a Dogecoin.
Total network capacity (currently, using BitInfoCharts http://bitinfocharts.com/dogecoin
): 84 Mh/s (Scrypt).
"beefy" GPU performance: 0.75 Kh/s.
"beefy" GPU cost: $450 + $75 (1/3rd of a rig, assuming 3 GPUs per mobo+power/supply rig)
So you would need 112,000 of those bad boys to get 51%. That's a $59 million investment in hardware.
But ..., here we have some very immature pools (in terms of infrastructure spending and experience) with a good chunk of the hashing capacity (scroll to bottom to see pie chart):
So you DDoS DogeHouse, Multipool and Fast-Pool, and now you only need maybe 40 Mh/s to be able to make your private Dogecoin blockchain fork to be the longest. I don't know if there still are many cheap GPUs available yet but at one time you could have bought or leased a ton of GPUs -- meaning acquiring 50,000 of them was possible.
But even this approach, with a $20M investment in hardware is not going to be a profitable attack, as you aren't going to extract $20M of value from Dogecoin through exchanges.
So today, Dogecoin (and Litecoin) appear to have a sufficient level of mining capacity as long as no attacker has invented and produced a Scrypt-based ASIC miner.
But let's say someone put in $4M of capital and has today 40 Mh/s of hashing capacity from the first week or two of production of the scrypt-based ASIC that party has produced. Can that party extract $4M of value from alt exchanges that support Dogecoin? Are there enough of these exchanges allowing anonymous accounts (or did poor enough KYC where a number of fake accounts exist?). If so, then it is entirely plausible that a 51% attack against Dogecoin (or Litecoin even) could be performed successfully, and done so at a profit.
Worse, depending on the jurisdiction that those exchanges operate, the pain could be felt even by those who didn't touch a Doegecoin or Litecoin even.
This is because in many jurisdictions, funds sent for deposit at a crypto-currency exchange are treated as liabilities owed by the exchange. It doesn't matter if the attack was against Dogecoin, if the exchange goes bankrupt (unable to honor all withdrawal requests) all accountholders are in the same boat. Your bitcoins at a Cryptsy are no safer than your Litecoins ... if Litecoin is successfully attacked.
So Bitcoiners who use alt-coin exchanges should be cautious with their funds left there. Kraken is the only exchange that has published their
plans on what would happen should an altcoin be attacked successfully with a 51% attack (or other failure that results in losses of customer's funds):