I think one of the reasons we got rejected was our response to some of Paul Graham's suggestions. One of them was to expand the concept to include game characters (Iconfinder is currently an icon search engine). We didn't respond to his idea in a positive way but instead defended icons as a large enough niche to build a initial service people would pay for.
I'm afraid that memleaks alone would be too small benefit to keep a dedicated overhead running in production. Leaks are something that you worry only after it already happened. What could work is a system monitoring dashboard that includes memory leak detection. Definitely something corp clients and developers could love at the same time. But the competition is already there, just none of them has the memleaks detection.
Zeroturnaround has been very successful with the java class reloading business thanks to the massive pain that java developers face daily with the bloated frameworks that make you develop inside the running application instead of writing small batches of test-driven code. Perfect fit for corporate environments. I'm not sarcasatic here. On the contrary, I'm very happy to see a very serious problem being solved. It's a huge dedication to work with all the containers and frameworks to get the live class reloading right.
Regaring VC funding and tools: VC's are surfing the latest waves to reap the profits and capture the market. Tooling is enabler, not a wave.
You can sell dev tools, but the market is small and fairly stingy. There are also lots of good free opensource alternatives, because your users are perfectly capable of building and maintaining their own tools.
[1] I believe the answer is probably Visual Studio, which as far as I know is run as a loss leader to promote Windows. Given the amount of development effort going into it, I doubt they sell enough copies to cover the cost.
Saying "met with top-tier VCs and angels" is usually good enough, without the downside.
