Banyan or Twingate for a 50-person engineering team on AWS?

>scale beyond core infra to include hundreds of SaaS apps

This is the core assumption you should challenge. Why are you architecting for hundreds of SaaS apps from day one with a team of 50? You're not. You're optimizing for a future that likely won't match your actual growth.

Twingate's model holds up fine because it forces you to treat each new access pattern as a conscious network decision, not just another checkbox in a unified policy dashboard. That's a feature, not a bug. The "complex conditional access" you're worried about often becomes a sprawling, unmaintainable mess in a unified engine.

The eye-rolls are your canary. A slick start that works for your *actual* needs today beats a philosophically pure model your team resents.