Rather flawed argument, when you consider that it's comparing general purpose servers to highly specialized asic/fpga miners. I have no doubt that if google re

Rather flawed argument, when you consider that it's comparing general purpose servers to highly specialized asic/fpga miners. I have no doubt that if google really wanted to, they could roll out the same kind of specialized mining hardware and dominate the blockchain, they just have no reason to. Consider, specialized hardware can get up to 1,000 MHash/s/$, which means an investment of $1000 would get you 1,000,000 MHash/s. Now let's look at google's datacenter expenditure. They spent something like $11 billion of datacenter expenses in 2014, and even considering that part of that is non-server related expenses(real estate and building infrastructure), let's see what happens when we punch $5 billion into hashing hardware. I get a value of around 5,000 PetaHashes/s, which would absolutely swamp the network, as it's only sitting at 35PH/s. Also, consider that that's just their expenditure for 1 year, not the total they've put into their datacenters. In conclusion, comparing cpus capable of general purpose computation to specialized hardware that can only do one thing is useless. After all, how many search querys per second can the blockchain handle?