Comment by nickpsecurity

The markets primarily want speed, compatibility, ease of view, and ease of deployment. They also want security if they can get that too. They will not pay for security in most cases. OSS developers work on what is fun to them which is usually the new design more than porting every bit of legacy code to the new design. Solutions usually must fit into these constraints.

Secure browsers cost a lot to develop. Developers needed to add security in a way that didn’t compromise any of the things that users carry about the most. Instead, they’d be making changes to browsers that ran on typical, operating systems. Expensive ports of browsers to separation kernels remain an academic curiosity.

The other issue is momentum. The highest assurance security community was focused on separation kernels that ran a combination of virtual machines and isolated apps. All open source and commercial efforts begin to run browsers inside VM’s’s. New projects followed suit. Here’s an example of that type of architecture:

https://genode-labs.com/publications/nizza-2005.pdf

Commercial examples are INTEGRITY-178B and LynxSecure. GenodeOS is a dual-licensed, OSS example.