True, the limit to an XML sitemap is 50MB (uncompressed) and 50,000 URLs. This is also the limit for any type of sitemap file, including HTML; any more than that, and the engine will stop crawling any single file. Also, @Evgeniy is correct that HTML sitemaps are pointless; nowhere in SEO literature does it say that Google or any search engine likes these types of sitemaps.
Sitemaps are a good way to hint to the engines which content is important, or has recently changed or been added, and submitting one in the Search Console is a great way to invite Google to re-crawl your site. (This is not a replacement for great website architecture and internal linking schemas.) Agree with @Stephen that sitemaps aren't nearly as crucial as they used to be, but in my own experience, they can still be helpful if done right. If you're going to have a large sitemap, best to stick with XML, since it's the standard.
Two things you can do here. You can have a single XML sitemap at the root, with only the most important and top level pages, plus newest content. The older content will be found and crawled if the site architecture is sound, or else if it was already indexed and no changes were made to the content, the engine won't even need to re-crawl it; it's already in.
For a more complete sitemap, you can break it up into multiple sitemaps. Basically, at the root, you'd create a sitemap index file, pointing to the URL's of all the sitemaps. You can break the sitemaps up by content type (pages, posts, etc.) or any way that makes sense for your site. A good resource here is Yoast; if you have a WordPress site, you might be using them already, but if not, look at how they structure their sitemaps and try to mimic that. If possible, it's best to automate sitemap generation and updates; otherwise you'll spend way too much time and labor on this.