I've done it myself also in the past because, like you, I didn't believe it either (this was around 2018 I'd guess).
It's a fairly simple calculation to take an energy-efficient graphics card and see how many SHA2 hashes per second you can get from that, see its energy consumption (ignore the rest of the computer and connectivity to support the GPU for now), multiply that with the hash rate, et voila you get some ridiculous number (don't remember, but something like half the world's energy consumption). This indicates that all Bitcoin miners are using ASICs. Now the question becomes how efficient a Bitcoin ASIC is, and when you hunt for that number and plug that in, the result is in the same range as what the news has been saying.
It also roughly works out when you look at kWh costs versus Bitcoin profit per consumed kWh, so the figure is not unreasonable. People would be leaving money on the table if their real electricity consumption were lower, because then they could add more equipment and make greater profits. Everything points towards this unbelievable waste to be real, unfortunately.
When you can be profitable running a datacenter without even needing to find customers to pay for your services on actual servers, why wouldn't you just run miners instead?
Clearly the incentive is there for this to be the reality.
It doesn't matter what you see driving down the street. There are widely available stats on how much energy mining takes (based on GPU energy usage, nobody's lying about this, you can run an energy meter yourself) and how much mining is done (verifiable with actual Bitcoin stats, this can't be falsified).
And obviously the warehouses aren't evenly distributed. They're concentrated where costs are cheapest (mainly energy costs).
What does GPU energy usage have to do with Bitcoin? Isn’t it predominantly mined by ASICs? (Though GPU and even CPU is possible, but uneconomical unless someone else is paying the bill...)
But I'll get you started: The estimate I find from various publications, both crypto friendly & otherwise, estimate around 1,500 kwh/bitcoin. Multiple out from there to the ~40% network hashrate capacity in the US and divide by total US kwh output.
Alternatively you can find a bitcoin profit calculator of choice and use its inputs to work backward.
Buy it or not. Your gut feeling doesn't have to correspond to the math, and I'm mostly okay with your gut being wrong (unless you're in a position to make policy decisions or something. Doubtful though...)
Take the network hash rate, the take the latest AntMiner S9's MH/J. Multiply it out. Weep.
That's gonna be your best case too, btw, assuming nothing less efficient is on the network.
[0] https://www.asicminervalue.com/miners/bitmain/antminer-s19-x...
Re the S19: it's not the hash rate that matters, it's the hash rate per unit energy. The S19 consumes 3X as much power as the S9 and yields 10X the hash rate so it's only 3.33X as efficient. Of course efficiency here is purely transitory as old miners are thrown into the rubbish bin and replaced with these new lotto scratching machines.
Why do you think your belief, unsupported by any facts, sources or calculations, has any value?
> Drive down your street,
Can you explain why this is some sort of reasonable way to find data centers? Why would you expect these data centers to be on streets where people live?
Drive down your street, look at all the businesses with their power on, all running computers. Every home running lights, fans, computers, video games, TVs, etc. then we have industrial manufacturing, etc.
You mean to tell me there’s some people in a warehouse around here with racks of servers running hashes? Just hashes, not a data center. We’d need hundreds of those warehouses to be 1% of the On-grid power. Sorry, I don’t buy it.