If your car runs on electricity, is taking your from place A to B waste of energy?
That energy is used to secure transactions. And even with its decentralized nature which is usually less efficient that centralized systems (but more reliable, and without single point of failure), it seems to be so far doing it more efficiently than banking systems.
Financial systems is one of those things we do because as human beings, we're handicapped in terms of coordinating with each other. We depend on external feedback loops to do anything as a big group. Perhaps if we were smarter, we could all just agree to do things for common benefit. But we're not that smart, and hence we need an economy. Maintaining it takes energy, but it should be seen as civilization upkeep costs - something we have to pay, but would do well to reduce as much as possible. It's a trade-off between amount of energy we use and the utility we get from it, and so it's fair to call unneccessary energy expenditure a waste.
The reason people call Bitcoin a waste of energy is that because it makes it clear how ridiculous the whole thing is - you're literally throwing away useful resources to generate abstract numbers that exist solely because we suck at coordinating with each other. In case of traditional banking the energy expenditure is hidden under many layers of indirection (the mentioned AC costs, for example), so it's not that obvious to people.
The other problem with Bitcoin is that resource usage is - as far as I understand the mechanics of that cryptocurrency - locked in a zero-sum game. You need to throw more and more electricity away just to keep up with other people doing exactly the same thing. This is the worst case of waste imaginable - because you can scale up your energy use ad infinitum without any marginal benefit. That's why also political campaign and advertising is mostly waste - everyone's effort just goes to cancel out the effort of everyone elses.
We'd do well as a civilization to reduce our upkeep to the minimum required.
(1) I am of course ignorant of peak loads here.
Not peek as in to look.
This is one of those annoying misspellings that is making the rounds recently, especially on reddit.
The task is not to have some spreadsheet with transactions that you could have of course on the oldest computers. The task is to provide a global payment system that people trust.
So obviously we're not going to use RPi. We need some solid machine. No problem, best server is cheap for the global payment system.
Well we also need reliability. So at least a few of these machines.
Real reliability, so actually they must be in some separate geographical locations.
At this point there are already some problems with consensus but they can be solved since we control the whole system. Cool.
But what if one location is compromised? We have to care about that, our global payment system must not be destroyed by servers being compromised in any way.
So first, we need some really good software security. Solvable, it can be open source, maybe people who depend on it will review it. Then we need hardware security. Security at these locations must be really good. It's global so we might actually want to be able to protect ourselves from some country's army - that bumps cost of these facilities a bit. Actually, a bit too much. We want to be independent from governments. They have a lot of money, it's really hard to protect physical locations.
Then there's the trust problem. Who runs it? Some organization I assume that have some cut of the cost of the whole operation. I won't even dive into details here. Throughout human history there probably was not a single organization that could be trusted, objective and rational once it gained enough power.
And to make it complete, unrelated to your comment, some people are mentioning proof of stake. Apart from the fact that it doesn't seem to provide any provable security, you have a trust problem again. Proof of stake is based on the assumption that most holders (in terms of money) want bright future for the currency. I wouldn't assume that. I'm happy that it is completely irrelevant for me who is Satoshi and what is she going to do with her coins.
And then, there's initial distribution problem...
I don't really see any more efficient solution to what Bitcoin currently provides. It is of course probable that such solution will appear in the future, but the network effect is huge. It's really first time in history we have such kind of money and Patrick Byrne (overstock CEO) explains it nicely in some of his talks. So even if such technology appears, I think it will get implemented in what is currently called Bitcoin (sidechains seem to be some baby steps of that).
(OK, I got a bit of topic, but these WASD keyboards are such a pleasure to type on.)
The price of Bitcoin is dictated by greed and speculation, not by its utility.