Having built (and debugged) large Next.js apps before (see my other articles series linked in the post), we applied those lessons learned to this decision. Factors like the friction in self-hosting Next.js at scale, the difficulty maintaining high performance with our specific page structure + third-party scripts, and the desire for faster dev cycles led us to explore alternatives. Astro's focus on minimal JS, Island Architecture, and framework flexibility ultimately aligned better with the specific problems we needed to solve this time around.
Full details on the comparison and decision process inside. Happy to discuss further!