diff --git a/CHANGELOG b/CHANGELOG
index 50ddd43..f9d2f91 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -3,6 +3,7 @@ version next
 - Fix "Underfull \hbox (badness 10000)" warning in banking \subsection dotted
   rules by adding \hss glue as in \dotfill (#307)
 - Update "circle" option to allow for half-filled circles in skill matrix (#309)
+- Fix "Arithmetic overflow" in cvcolumns when every column sets a width (#311)
 
 version 2.6.1 (24 Jun 2026)
 - Allow the customisation of the contemporary style (#297)
diff --git a/moderncv.cls b/moderncv.cls
index 1bc93cf..1f02b2a 100644
--- a/moderncv.cls
+++ b/moderncv.cls
@@ -676,7 +676,9 @@
     % compute the width of each cvcolumn, considering a spacing of \separatorcolumnwidth and the columns with set width
     \ifnum\thecvcolumnscounter>0%
       \setlength{\cvcolumnswidth}{\maincolumnwidth-\value{cvcolumnscounter}\separatorcolumnwidth+\separatorcolumnwidth}%
-      \setlength{\cvcolumnautowidth}{\cvcolumnswidth*\ratio{\cvcolumnsautowidth}{\cvcolumnsdummywidth}/\value{cvcolumnsautowidthcounter}}\fi%
+      % only share the remaining width if some columns have no explicit width, to avoid a division by zero ("Arithmetic overflow") when all have one
+      \ifnum\value{cvcolumnsautowidthcounter}>0%
+        \setlength{\cvcolumnautowidth}{\cvcolumnswidth*\ratio{\cvcolumnsautowidth}{\cvcolumnsdummywidth}/\value{cvcolumnsautowidthcounter}}\fi\fi%
     % pre-aggregate the tabular definition, heading and content (required before creating the tabular, as the tabular environment doesn't like loops --- probably because "&" generates a \endgroup)
     % - the tabular definition is the aggregation of the different "\cvcolumn<i>@def" (by default "p{\cvcolumnautowidth}"), separated by "@{\hspace*{\separatorcolumnwidth}}"
     % - the tabular heading is the aggregation of the different "\cvcolumn<i>@head", separated by "&"