In the Bitcoin world, the term "premining" is a dirty word and for a good reason. There have been numerous preminted altcoins out there created with the sole purpose of being pumped and then dumped. Some other premined coins gain slightly more credibility by premining their tokens in order to create fiat-denominated cryptos. Then again, we had some big scams in that realm as well.
Over the last few years we also had a few token presales for various app-coins (Augur, MaidSafe, etc.) and platform tokens (Ethereum, Mastercoin, etc.). While the tokens have mostly been advertised as being sold to give you access to the platform / application (possibly to avoid securities regulations), it is clear that a lot of people purchase those tokens in hopes of speculating on their future price.
And lastly, we have the case of Ripple, a Crypto 2.0 platform that was completely premined and whose parent company still owns a vast majority of the XRP tokens. Some have been sold, some have been given away, but the fact still remains - Ripple Labs owns most of those tokens. Some maintain that it can be a long-term business strategy for the company - to hold onto the XRPs for a long time while building up the network to earn money from the appreciated value of the tokens. While this might be a useful stream of revenue, I think there might be a better use for such tokens for both Ripple and similar systems that might come along.
As we have seen recently, a lot of big institutions are waking up to the idea of "a blockchain" and its usefulness for accountability and so on. Ripple is among one of such blockchains pushing for being adopted by big companies like banks or Western Union.
However, from what I heard chatting with some people in the industry talking to such institutions, the hard sell in a lot of cases is the token / coin. Perhaps this is why we see banks not wanting to adopt Bitcoin but being enamoured with a more bland and generic "blockchain" - you need bitcoins to use Bitcoin, and the price of bitcoins fluctuates wildly. The institutions don't want to get exposed to the price swings, so they opt away from such platform. However, what if we could remove that uncertainty altogether? Well, in some cases we might just be able to...
Imagine we have a company that wants to get onto a blockchain and is weighing its options. One of their concerns would be whether a system they build today will hold up in a few years time, both in therms of technical capabilities (will the system be able to expand to meet the growing needs of the network, or will we have another block size debate) and pricing (will the costs stay the same or go up). While the earlier is always up in the air since you can't predict everything, the latter might have a more concrete solution.
If we took a company like Ripple Labs with their big supply of XRPs, they could possibly offer a good solution to the pricing problem for an important enough customer. All they would simply need to do is offer that company a long-term option to buy the tokens at some fixed price. This way, the company could be certain they won't pay more than X to use the system in the future, no matter where the price of XRPs might go in the future. Whatever Ripple Labs might be losing by entering the option contract and possibly selling a fair amount of tokens at a low price years down the line, it could make up in other areas - either integration fees or the rest of their XRPs appreciating in value due to the high profile of customer they brought on.
So all in all, while there are many reasons why premining is a bad practice and should generally be avoided, there are a few reasons why it might be useful for the growth of the system.