While I don't think Apple intends for undersupply per se, I think that in the doubt, they guess low very intentionally.
As many have pointed out, they have a history of doing this.
There are a couple of reasons this makes sense for them.
1. If you get to little supply, it leads to positive press and buzz about the product.
Remember the ipod, and the ipod mini, etc.? Every week in the tech, business, and sometimes front section, you'd get a story about people going to great lengths to get one for christmas, with this insane demand. Apple is about buzz, and a shortage helps create it, while a surplus kills it.
2. If you get too much supply, you've got to cut prices to get rid of product.
Apple is loathe to do this - it says that the product is not worth as much as apple says it is, and perhaps that the whole brand is in the same boat. Apple is about brand, and having to work to get rid of oversupply really hurts the brand.
How many times will apple wait forever to update a product so that they can let supplies of the old one dwindle, only to release the new and better one at a lower price? Clearly the old one doesn't cost more to make, and it isn't helping apple to have an out of date product, either. But if they cut the price to move it, you see failure there, you see having paid too much for it in the first place. If they wait and make the next one cheaper, they get better buzz about the new version, and you see the incredible pace of apple's innovation making you want the new one, rather than having paid too much for the old one.
3. Apple has to make wild guesses about supply often. They bet big on products with big unknowns as a course of business.
So, apple has learned to do low risk, high hype, high brand building behavior here. Other companies don't really need to do this, because they aren't running like apple is. Most companies sell a ton of different products, innovate rarely (preferring to buy or imitate smaller companies who have proved an idea successful), and when they do make something bold and new, they start small and let the product grow. Even google, insanely innovative, opts for a proliferation of interesting side projects that don't take up that many resources, and then watches to see what grows and needs more support.
Not apple. While apple's bets are usually safer than they appear (they think things through) and they very intentionally sit on a very large pile of cash, should things ever go bad, it's still huge. Let's assume that they make commitments for about a year's supply at a time. roughly, I'm calling iphone's first year sales 4-5 million, and it looks like ipad is already well ahead of that, (without the wifi version). So let's say we see sales of 10 million the first year, which I feel is conservative.
Now, look at the ipad. It's unprecedented. It's unnecessary. It's unbelievable. Sales are totally unknown, and nobody should be surprised if apple sells double or half of their predictions. Including apple.
So if apple overestimated by double, they'd have an extra 10 million ordered, which at 500 bucks a pop (I'm very roughly balancing out the variety in models sold with apple's typical 30% margins and saying, let's price it at the cheapest model) is a FIVE BILLION DOLLAR LIABILITY. You can argue with numbers here, and about the pricing model, etc., but you'll still end up with what, a 2 BILLION dollar liability?
Yikes.
Now, underestimate by half, and you're potentially missing out on 2.5 Billion in revenue (I know, I know, I'm mixing revenue and costs and margins messily, but I'm lazy, and this is getting long). But you're still getting 2.5 bil, while giving the brand huge cachet and earning a very healthy margin. Adjust the next version's volume accordingly, and you just might be able to cash in on all the hype and anticipation enough to make up for the money you left the first time around.