Here's my 2 bitcents:
The influx of new users drives the exchange rate.
The exchange rate drives the mining power, with some delay.
For 2 Th/sec we currently need about 1MW electricity. With these costs, the system cannot grow on speculation alone, for very long.
When the user growth slows, current miners will slowly bring down the prices, even if no early adopters cash out.
From the demand side, there's still lots of growth potential.
$10/BTC is psychological resistance and we'll probably retest the lower single digits. If $20 is breached, we could see $100 at the end of the year.
Maybe Bitcoin really does become as big as PayPal, then these prices are justified, but imho the chances are overestimated.
Users stop to grow, miners then slowly push prices down, other miners stop to mine because prices go down, with less miners difficulty goes down, with lower difficulty miners earn more bitcoins and then dump the market, and so on....