Risky...
Risky.
The upside: Hypervisor software is relatively small. This isn't like writing an entire OS, or an office suite, or a web browser. Also,if your business is based on "cloud" or "convergence" or whatever (i.e. hypervisors), then rolling your own DEFINITELY lets you differentiate your software and services from the other providers in the market. As AC says, being able to spin up both VMWare and Hyper-V VMs is pretty unique already. This is really a rather immature market, and there's no real guarantee that any of the existing hypervisors are particularly close to an "ideal design" yet, they may be able to come up with a noteably better design.
The downside: The software, although small, is highly technical. It could be entirely possible to have development hit a snag, or end up with a hypervisor where the major bugs aren't quite worked out (which of course makes it quite the non-starter for the types of uses hypervisors are used for.) It's possible for you to think some features are important, but find the customers don't think so. Finally, it's possible to come out with some new features but have the other vendors replicate them.
