[amp-email] Could runtime loading and sandbox initialization affect email open rates

[johninch]

Title: [amp-email] Could runtime loading and sandbox initialization affect email open rates?

Background

We are investigating a significant decrease in email open rates after introducing AMP Email.

Our AMP emails use components such as amp-carousel. When the AMP version is rendered, we can observe additional resources being loaded, including v0.js and amp-carousel-0.1.js.

Questions

We would like to understand whether the following AMP Email initialization steps could introduce noticeable latency:

1. Runtime caching
   Are v0.js and component JS resources cached and reused across different AMP emails, or can they be downloaded again when opening each email?

2. Sandbox initialization
   Does the AMP Email rendering process require creating a separate sandbox environment? If so, is there any known optimization or ongoing work to reduce the initialization overhead?

3. Performance impact
   Are there recommended approaches for AMP Email senders to minimize runtime loading and sandbox initialization latency, particularly for components such as amp-carousel?

Any technical insight, documentation, or recommendations for measuring and debugging this behavior would be greatly appreciated.

Thanks!