You Only Need to Include Standalone AMP Pages in Sitemaps
As long as you have the link rel=amphtml you won’t need to include AMP in sitemaps unless they are standalone pages. Google can access the HTML of the main page to include in the AMP cache when changes are made to the content.
Google Crawls AMP Pages to Validate & Cache Them
Google crawls AMP pages to check they are valid with a canonical tag back to the webpage, and also to be able to show them in the AMP cache.
Some Features in Organic Search Require AMP For Security Reasons
A number of search features require AMP to work well e.g. news carousel. For these search features it isn’t enough to have a mobile-friendly website as Google can’t serve your content from Google.com’s cache for security reasons.
Implement Tracking That Joins AMP & Normal Page Sessions
There are solutions to implement analytics tracking on AMP which don’t increase bounce rate, so that sessions are matched together when a visitor goes from the AMP version of a page to the normal version.
AMP Pages Need to Have Canonical Tags
AMP pages without a canonical tag run the risk of not being considered to be valid, as well as creating duplicate content issues.
AMP Can be Used For Desktop Pages
You can make desktop-friendly sites using AMP because it is a responsive website design framework which can be used for different types of content and pages.
AMP is Not a Ranking Signal
Rankings will not be altered by having AMP pages or not.
AMP Pages With Reduced Content Won’t Be Treated As AMP Pages
AMP pages with less content than on the mobile page or desktop page equivalent won’t be treated as AMP pages by Google. Users also shouldn’t have to click to read more content from AMP pages.
Old AMP Pages Can be Redirected to New AMP Pages
Old AMP pages should be redirected to new AMP pages. The AMP cache will try to update AMP and if these pages are found to 404 then they are likely to be dropped from the AMP cache rather than Google working out that they have moved to a new URL.