Here’s some of the reasons, in no particular order:

• Kernest delivers a number of open-sourced fonts, seems only appropriate to open source Kernest’s font serving engine.
• Almost all the software I use to create, manage, and maintain Kernest is open source. Again, seems only appropriate to open source Kernest’s font serving engine.
• It’s the easiest possible way to get Kernest where my client’s prefer it – in their infrastructure.
• Kernest shouldn’t be the bottleneck to @font-face adoption.
• The most efficient ways to implement @font-face are continually changing. Open source is a way to distribute these learnings across the community as quickly as possible.
• I want the code to be better.
• It gets Kernest closer to a Affero General Public License, something I’m working towards and I think more software-as-a-service offerings should.