No don't think it violate the Mises regression theorem, it is just that it evolved incredibly fast so the first stages where barely noticeable.
BitCoin was not actually created as a money, but a scarce virtual good that can be owned anonymously. The first transactions with BitCoin would have been barter and not indirect exchanges. People mined some coin and used it to barter to something else, and only gradually there are more and more indirect exchanges where BitCoin is a money taking place.
The fact that a owner of a unit of BitCoin good can not be directly identified and that it is virtual add tremendous utility value to BitCoin. Also a bit of the initial value can be derived from the nuseanse and cost of creating it. Mises theorem is simply and understandably just not adapted to a virtual economy so it is difficult to see, but there is utility value in BitCoin that is not purely monetary. It don't think Mises ever claimed something first have to be used in heavy industry, gold didn't have any industrial value before it was used as money. It was just pretty to make ornaments from and scarce. With BitCoin instead of wanting shiny things to give the money initial value people wanted something which is virtual and who's owner is very difficult to identify...
BitCoin was developed as a currency from the begining. Just look at its name bit
coin, and the paper that originated them "bitcoins: a peer to peer
cash system".
Transaction privacy is a monetary utility. Costs are irrelevant for market value, I could spend lots of energy and work to produce something that nobody wants, so no matter the cost its market value will be 0. Market values drive costs, not the opposite.
When I talk about Industrial use I mean any non monetary use, which includes ornamental uses. Jewerly is a pretty strong industry. Gold had been used for ornamental purpose well before it was used as currency. Mises claims that for a good to became money it needs to have a non monetary prior use, this is what his Regression Theorem is all about.
But I claim that Regression Theorem is not necessary, and it does not work not only with BitCoins, it doesn´t explain well why credit currencies have value. Carlos Bondone´s monetary theory, which is based on Austrian School founder Carl Menger,proposes a stronger monetary theory, which prefectly explains the BitCoin phenomeon and in regard of this theory
BitCoins do qualify as money.