Actually the projected time for all coins to be mined out is approximately 130(ish) years.
The official time is that the coins should be all mined out around 2140, but as the hashrate for a majority of time will be higher than the avg rate the previous 14 days or the difficulty will be raising most of the time, therefore there will be a slight overproduction of coins. 2016 blocks will be produced faster than every 14 days on avg so it will probably take quite a bit less than 27 years to produce all the remaining coins, so maby sometime between 2035-2037 or so there will be no more coins to mine.
That will not happen. It might be maybe 1-3 years before, with the current difficulty adjustment it would not be practical for all the coins to be mined 8x faster?
I think you're both right, although TF is right for better reasons.
The block reward will almost certainly halve beneath 1 satoshi / block (this being the definition of the end of the block reward) before the standard projected dates. However, we are still very early in the ASIC era and are seeing the sorts of exponential changes in network hashrate / difficulty that are absolutely impossible to maintain. The hashrate will continue to grow parabolically only until the network is saturated with ASICs - at which point the gains will become incremental as we just add new hardware of the same type onto the network. Generational upgrades to the ASIC infrastructure will provide a sort of leg-up to the hash-rate only occasionally, and with diminishing returns.
Jumping from 128 to 65nm, for example, will improve GH/J but probably wouldn't improve GH/btc on the initial purchase point (at least) - and we're absolutely nowhere near knowing the life expectancy of these chips, but a cursory glance at relevant journal articles such as
this one would seem to indicate that sub-130nm chips have a much shorter life expectancy than what people might be thinking. This would be a much better business case for hardware sales than mining, to be sure.
Growth in network hash-rate is going to get off the parabolic rocketship at some point and flatten out into slower growth and incremental improvements. Once we reach this point we will stop seeing 6-8minute average time to block for a difficulty period and start getting closer to 8-10 minutes, I expect. This may not happen for quite a while, of course, at least from the perspective of most in this community.
This is all somewhat irrelevant, however, which is why
thy was still somewhat correct.
While the block reward will last for quite a while yet, the real question is how long will it remain meaningful.
75% of all Bitcoin that will ever exist will exist at the time of the second (next) halving. 87.5% of all Bitcoin that will ever exist will exist at the time of the third halving. This process will only continue, and we will have issued >99% of all Bitcoin that will ever exist at the time of the 6th halving sometime before 2033.
What some have observed is that, regardless of this, all coins that actually are in circulation will continue to cycle through the blockchain forever - with some small % of them being paid to miner's as fees. Many have previously noted that the fees will need to out-weigh the block reward itself eventually, however most people tend to talk about this as 'what happens after the end of the block reward'. This will actually need to happen much sooner than that.
At the moment the average fee per block is, interestingly, almost exactly 0.1 BTC - I get this by dividing the trailing 24-hour fees (19.24 BTC) by the number of blocks (192) on blockchain.info. At this rate, even with no change in fees the block reward would drop below the current fee at the 9th halving sometime before 2044.
I guess my only real point is that the minting of new coins is mostly irrelevant to the long-term prospects of miners.
note: thy added to the discussion since I started writing this but I can't really be bothered to edit.
