diff --git a/include/fmt/format.h b/include/fmt/format.h
index e6c649d2..8a4ab7fb 100644
--- a/include/fmt/format.h
+++ b/include/fmt/format.h
@@ -1715,7 +1715,12 @@ template <typename Range> class basic_writer {
       return;
     }
     int precision = specs.precision >= 0 || !specs.type ? specs.precision : 6;
-    if (fspecs.format == float_format::exp) ++precision;
+    if (fspecs.format == float_format::exp) {
+      if (precision == max_value<int>())
+        FMT_THROW(format_error("number is too big"));
+      else
+        ++precision;
+    }
     if (const_check(std::is_same<T, float>())) fspecs.binary32 = true;
     fspecs.use_grisu = use_grisu<T>();
     if (const_check(FMT_DEPRECATED_PERCENT) && fspecs.percent) value *= 100;
diff --git a/test/format-test.cc b/test/format-test.cc
index 42ba77ba..7f718e7a 100644
--- a/test/format-test.cc
+++ b/test/format-test.cc
@@ -1236,6 +1236,9 @@ TEST(FormatterTest, Precision) {
   EXPECT_THROW_MSG(format("{0:.2f}", reinterpret_cast<void*>(0xcafe)),
                    format_error,
                    "precision not allowed for this argument type");
+  EXPECT_THROW_MSG(format("{:.{}e}", 42.0, fmt::internal::max_value<int>()),
+                   format_error,
+                   "number is too big");
 
   EXPECT_EQ("st", format("{0:.2}", "str"));
 }