Theorem 5.2.5 (Descent for colimits). Let \(I\) be a small \(\infty \)-category, let \(\overline {F}, \overline {G} \colon I^{\triangleright } \to \An \) be functors and let \(\overline {\alpha }\colon \overline {F} \Rightarrow \overline {G}\) be a natural transformation such that the restriction \(\alpha := \overline {\alpha }\vert _I \colon F \Rightarrow G\) is a cartesian transformation, in the sense that for every morphism \(i \to j\) in \(I\) the commutative square

Commutative square with maps F(i) to F(j), G(i) to G(j), F(i) to G(i) labeled alpha(i), F(j) to G(j) labeled alpha(j).

is a pullback square. Assume that \(\overline {G}\) is a colimit diagram. Then \(\overline {F}\) is a colimit diagram if and only if \(\overline {\alpha }\) is a cartesian transformation, i.e.Β also all the squares \begin {equation*}

Commutative square with maps F(i) to overlineF(infinity) overlinealpha(infinity), G(i) to overlineG(infinity) = \colim I G, F(i) to G(i) labeled alpha(i), overlineF(infinity) overlinealpha(infinity) to overlineG(infinity) = \colim I G.
\end {equation*} are pullback squares.

Proof. See Section 24.7; see also [Lurie (2009b), Theorem 6.1.3.9]. β–‘