So far for the projects that I'm involved in I have yet to find a single case where I can make a case for a cloud solution vs hosting the stuff directly. (but there are other reasons than financial ones to use a cloud).
Where are you located ?
Or are you really burning through SQS requests fast enough to make it cost cash money?
Either way, it sounds like you've painted yourself into a corner it will be hard to get out of. AppEngine is probably the most difficult platform imaginable in terms of how hard it is to migrate off of. You'll have a ton of work to do just to get out of BigTable, and then you're left with a Django app.
I guess the best you can do is start charging money, and start developing a new version as quickly as possible to stop the bleeding.
Good luck!
We released on GAE and we're also using Amazon SQS. At our peak load, we were getting over 1.5-2k users streaming in during the initial price drop. CPU Time was our biggest drain: we've combined instant messaging with a local and global search option which eats up resources fairly quickly. I was very surprised at GAE server costs given a userbase that is <10k at this point (given the costs I had to put the breaks on growth very quickly).
The VC approach was for seed-stage capital (~100k) - I found a few "programs"/funds targeting early stage companies. Basically, I'm looking for a business angel that could assume a mentorship role and provide some "smart capital" at this point to scale.