Choosing between React Router v6’s lazy loading and manual route preloading for bundle size vs latency

Goal: determine which approach yields optimal trade‑off between initial bundle size and navigation latency given variable network conditions.

Built‑in lazy loading reduces payload but may show Suspense fallback on slow connections; manual preloading keeps payload small while fetching ahead, eliminating fallback but increases request count and code complexity; memory usage similar unless many routes preloaded.

Uncertainty: how network speed thresholds affect perceptible fallback duration, and whether the added request overhead of preloading outweighs latency gains on typical user connections.

On a typical 3G connection, does manual preloading noticeably reduce perceived navigation delay compared to built‑in lazy loading?

Does the extra request overhead of preloading offset its latency benefits on broadband?

At what number of concurrently preloaded routes does memory usage become a concern?