My React Server Components (RSC) mental model is basically: They’re getServerSideProps / getStaticProps but more powerful. Things RSC / gSSP / gSP share in common: 🔸 You can’t pass things like normal functions from RSC / gSSP / gSP to the client, because they just can’t be serialized; 🔸 They must be executed before Client Components; 🔸 They don’t run in the browser; 🔸 They don’t have states, so no hooks; 🔸 They can both be executed and stored as static at build time (RSC / gSP), so they don’t require a server. Things that are more advanced in RSC compared to gSSP / gSP: 🔹 Server Components can pass elements and more to the client, instead of being limited to JSON values; 🔹 Server Components are able to define extra wrappers around Client Components, whereas gSSP / gSP can only affect components’ inner content; 🔹 You can have many Server Components and they will be more parallelized when executing, instead of only one giant gSSP / gSP blocking on the top; 🔹 Individual Server Component boundaries can be optimized as static if possible, instead of picking between fully dynamic (gSSP) and fully static (gSP) for all data; 🔹 Individual Server Component boundaries can be streamed to the browser and rendered earlier, unlike gSSP that must block the initial client rendering. – at The Hague, The Netherlands