First, the Pandora, like the GP2x and the GP32 before it, are not "mass-market" items. Nintendo might be able to sell several million units in a few weeks, Pandora won't. How many people do you personally know in real life that have even heard of the Pandora (other than through you)? How many members are there on this forum, or every gamepark-based forum? What percentage are likely to buy from the "first batch" (I know I wouldn't, until someone has posted a real review of how good it is and a few thousand people have tested them) or even at all?
Secondly, there's absolutely no sense in producing more than you know you can sell. The only way to know what sells tends to be "make some and see if you can sell them". So, making anything more than 3000 units and you'd be risking going bankrupt before you've sold a unit, or getting stuck with thousands of unsold (but with manufacturing paid for) consoles.
Third, investors tend to like to see how something succeeds, not sustain it indefinitely. So the 3000 is also a "taster" to make sure it can sell and make money before you get into needing serious money to ramp up production. Plus, you can pick up on quality control issues, make design changes etc.
Fourth, if these items are being mass-produced, you'd probably find that in units less than 3000 the manufacturers won't even entertain the idea and in units much more than that you can't afford to do it on the basis of "expected sales". They want to be able to build production lines, train staff, etc. to build thousands at a time, not ones and twos.
Fifth, there is *nothing* to stop the designers ordering more units from the manufacturers (probably in batches of thousands, and probably a bit cheaper than the first batch) once they have been made once and sold enough. The money from the first batch pays for the second batch to be made, the money from the second batch pays for the third and so on. Each step you make profit.
Sixth, say there IS a problem with the manufacturing. You just ordered 3000 consoles that are somehow faulty. Maybe you can fix them with a 1p resistor and a bit of soldering but better to find out on the initial 3000 than 10,000 or 50,000.
This is before you even get into availability of components (e.g. LCD screens), capability of the manufacturers (what if they can't make more than 1000 a month because of other jobs etc.?), packaging, shipping, talking to retailers, distributing samples to reviewers, distributing samples to developers and the million and one other things that have to be paid for.
Think about it. Say there's a significant markup on the Pandora, so each one costs only £100 to produce and sells for £300 - that's £300,000 for manufacturing the initial batch of 3000 before you've sold anything, before you've paid anybody, before you've packaged, shipped to a distributor, distributed them to retailers, before you even know that it works when it's been mass-produced... that's a lot of money.
The owners will make it back in the future *IF* Pandora sells, and some more. But say it's a sell-out and they get £300 per unit... that's £900,000 income - out of which you'd need to be paying AT LEAST another £300,000 to manufacture ANOTHER 3000, plus a significant amount of the boring overheads. And more likely, you'd be looking at trying to produce 5000-7000 of them the second time around, so your spare cash is in a constant cycle of "we made money, we have to spend it to make some more". And if you mess up the cycle (by mis-predicting demand), you can easily come out without any money at all and an awful lot of unsellable stock on your hands.
It's a risky business, which is why investors are usually very rich and can afford to take the chance. On the whole, you *usually* come out a winner. But if you're playing with someone else's money (i.e. an investor has said they'll pay for you to produce thousands of handheld consoles on the basis that he gets 50% of the profit), then you can't just blow all of it on a mass production run. You have to tread carefully and make money at each step, even if it's just enough to pay the bills and produce some more products.
Disclaimer: I know nothing about mass production of handheld computer consoles.