My argument against writing the browser for the Xbox is that it's more trouble than it's worth. There's even a stronger argument against a VM. Writing a Java VM from scratch is probably tougher than you're aware... study VMs a bit, and you'll see why. Finding an open-source VM for another OS will provide minimal help, since the whole point of a VM is to create an abstraction layer that's runs cross-platform code, and thus is very dependent on OS specifics.
Trust me guys, I don't shoot down these browser ideas because I think it's fun. If you want to try it, I'd love to see some attempts made at it - and I'd have alot of respect for anyone who got something going. I'm just letting you know why you haven't seen it done yet.
If you're a programmer posting these ideas, why not make an attempt at it? If it interests you, go for it! That's why all of us are here... to do what we enjoy!
EDIT: Dreamcast also used a windows CE-based OS, yeah? So it's a completely different task than writing a DX-based browser. And most importantly, the xbox project has no financial support (as someone already pointed out)