By using Compute@Edge to pre-warm the cache, you are not only using a powerful, globally distributed network to do the work, but you also solve the pitfalls associated with legacy prefetch.