Only by the addressable memory on each node, which varies. It would take a phenomenal number of unconfirmed tx to break nodes, but maybe enough of them would happen if wallets keep sending tx for thousands of frustrated BTC holders who are trying to get coins onto the exchanges as the price falls way below any TA support.


Oleganza is nuts if he thinks that $100 tx fees would ever be normal. How could Bitcoin keep any market share in that situation if another crypto-coin had 10x the volume and fees of about $1? How could Bitcoin keep any semblance of "digital gold" in that environment? (leaving aside the definition of it).

There is another reason why Bitcoin should not be allowed to hit a network-wide hard limit at any time, and ironically, for a trust-less system, it is trust in the system as a whole.
All users of Bitcoin have to trust that the trust-less network works and is viable. It has proved this over 6.5 years. That trust is hard-earned. Think about a relationship or a business partner, it takes years to earn trust, but that trust can be squandered over one serious event. If a partner cheats after years of loyalty, then how many years does it take to regain trust in that person again? Never?
Same with bitcoin, because: money.
A loss of trust just once is a scar for life.