If you're raising capital for a technical startup, how much should you take? Marc Andreessen says as much as possible, a sentiment echoed by Jason Calacanis. Jeremy Liew expounded on some of the risks of having a valuation that's either too high or too low. Marc and Jason are two guys who definitely know about building successful companies (and, perhaps more importantly, successful exits).
Jeremy's example focused on an angel round that gave a company a valuation of $30m, where he most certainly would have valued the company less than that (and thus either passed on, or lost, the financing). Despite what Marc and Jason say, though, you can raise too much money in a given round - not because of subsequent rounds, but because of the expected exit. After all, as Jason says, if you raise "too much" money in an earlier round, the big challenge is just not too burn it too fast.
This applies especially to VC rounds, which are typically very standard as compared to an angel round. The "average" fund has a life span of 10 years - the first 2 may be spent primarily raising funds, while investments are made typically up to year 4. That means that most investments need to be liquid in 5-6 years. The important thing to remember is that VCs invest for capital appreciation, not income - they are not interested in a cash cow, but rather in something that will "blow up". A good general rule is that the VC is looking for a return that will double in value and be marketable within 3 years.
If you take $10 million in an A round (giving the "standard" 40% per round), your company is valued at $25 million pre-money. That means that the VC is going to be looking for an exit within 6 years (preferably sooner) for at least $50 million. In many situations that might be fine - but there are also plenty of companies who might be very, very successful and profitable but not worth that much in the short term. Or, if they are valued accordingly, it still might undervalue what the company would be worth in the long run. (See, e.g., the number of Web 1.0 startups who bought back stock, etc. and eventually sold for significantly more than they would have been worth at the end of the VC window).
The real problem with raising "too much" money is that it creates certain expectations for an exit that might not be in the company's, or the founders', best interests.
(Edited to clarify a few points in the original post)
Update: After I posted this, Dick Costolo wrote an absolutely amazing post about raising too much money.