What it is
On a closed manifold, Ricci flow exists for a short time from any metric. Hamilton (1982) showed that it continues as long as the curvature stays bounded, so a finite maximal time Tmax means curvature blows up. Sharper extension criteria show which weaker quantity must already blow up. Šešum proved that a bound on Ricci curvature is enough. Wang showed that, with Ricci curvature bounded below, an integral bound on scalar curvature suffices. In dimension three the Hamilton–Ivey estimate shows that a bound on scalar curvature suffices; in general dimensions this is open.
When the flow exists for all time it is called immortal, and the question becomes what it converges to.
- Convergence to canonical metrics. Hamilton's first theorem: on a closed 3-manifold with positive Ricci curvature, the normalized flow converges to a round metric. On surfaces the normalized flow converges to a constant-curvature metric (Hamilton; Chow). Cao proved that the Kähler–Ricci flow converges to the Kähler–Einstein metric when c1≤0.
- Collapsing and geometric decompositions. Hamilton's non-singular solutions, Lott's analysis of Type III solutions, and Bamler's proof that the flow with surgery on a 3-manifold needs only finitely many surgeries and then has curvature decaying like 1/t, show how the geometric pieces of geometrization appear in the long-time limit. Böhm and Lafuente proved that blow-downs of immortal homogeneous flows converge to expanding solitons.
- Extrinsic flows. Ecker and Huisken proved that entire graphs evolving by mean curvature exist for all time and, under a growth condition at infinity, become self-similar expanders.
Why it matters
Long-time results turn the flow into a classification tool: if the flow exists forever and converges, the manifold carries the limiting geometry. Extension criteria are how one proves that a singularity of a given type cannot form. And convergence rates, often exponential near a stable limit, underpin the stability theory of Einstein metrics and solitons.
Where it stands
Open questions include whether a scalar curvature bound always prevents singularities (Bamler showed that flows with bounded scalar curvature converge smoothly away from a set of codimension four), the long-time behaviour of immortal flows in dimension four and above, and convergence of the Kähler–Ricci flow on varieties that are not Fano or of general type (see analytic minimal model program).