Geez guys, it's easy. I don't get what there is to not understand :
1. These launches result in a massive demand burst over the first few days. Demand then goes down to normal levels. This is predictable, Apple will sell phones very quickly in the first few days.
2. If you have a possibility of making 12 million phones for launch, anticpate a burst of 8 million to cover pre-orders and initial launch week sales and know you'll sell the other 4 millions over the course of the next 2 weeks easily then you can skip to 3 :
3. Limit initial offering to 6 million phones. Make sure to allocate a low number to initial pre-orders, around 1 million or so units so that the website quickly shows "Available in 1-2 weeks". This creates scarcity. Send the remaining 5 million to stores, making sure that your big markets get barely enough to cover demand. The "website said they ran out!" scare is in effect, the line-ups will form to get that coveted device first day since it's now impossible online.
4. Launch day, sell out of the remaining 5 million, making sure to leave a few sad faces for the cameras. Trickle extra inventory from your extra 6 million into stores each day, feeding the people who failed to get it on launch day.
5. After 2 weeks, you've sold the same 12 million phones to the same 12 million people had you offered the 12 million units up front. You've also probably managed to sell an extra 2-3 million off the media campaign covering the line-ups and general insanity.
You've lost nothing. You've gained free publicity. Heck, I'd do it if I had a business. No one is saying to sit on inventory long term, just not to dump it all into the market in one go.