diff --git a/clippy_lints/src/float_literal.rs b/clippy_lints/src/float_literal.rs
index c51267567d01..56f0873c1455 100644
--- a/clippy_lints/src/float_literal.rs
+++ b/clippy_lints/src/float_literal.rs
@@ -13,6 +13,8 @@ declare_clippy_lint! {
     /// Checks for float literals with a precision greater
     /// than that supported by the underlying type.
     ///
+    /// The lint is suppressed for literals with over 40 digits.
+    ///
     /// ### Why is this bad?
     /// Rust will truncate the literal silently.
     ///
@@ -126,13 +128,23 @@ impl<'tcx> LateLintPass<'tcx> for FloatLiteral {
                         },
                     );
                 }
-            } else if digits > max as usize && count_digits(&float_str) < count_digits(sym_str) {
+            } else if digits > max as usize && count_digits(&float_str) < digits {
+                if digits >= 40 && is_const_item(cx, expr) {
+                    // If a big enough number of digits is specified and it's a constant
+                    // we assume the user is definining a constant, and excessive precision is ok
+                    return;
+                }
                 span_lint_and_then(
                     cx,
                     EXCESSIVE_PRECISION,
                     expr.span,
                     "float has excessive precision",
                     |diag| {
+                        if digits >= 40
+                            && let Some(let_stmt) = maybe_let_stmt(cx, expr)
+                        {
+                            diag.span_note(let_stmt.span, "consider making it a `const` item");
+                        }
                         diag.span_suggestion_verbose(
                             expr.span,
                             "consider changing the type or truncating it to",
@@ -196,3 +208,19 @@ impl FloatFormat {
         }
     }
 }
+
+fn is_const_item(cx: &LateContext<'_>, expr: &hir::Expr<'_>) -> bool {
+    let parent = cx.tcx.parent_hir_node(expr.hir_id);
+    if let hir::Node::Item(itm) = parent {
+        return matches!(itm.kind, hir::ItemKind::Const(..));
+    }
+    false
+}
+
+fn maybe_let_stmt<'a>(cx: &LateContext<'a>, expr: &hir::Expr<'_>) -> Option<&'a hir::LetStmt<'a>> {
+    let parent = cx.tcx.parent_hir_node(expr.hir_id);
+    match parent {
+        hir::Node::LetStmt(let_stmt) => Some(let_stmt),
+        _ => None,
+    }
+}
diff --git a/tests/ui/excessive_precision.fixed b/tests/ui/excessive_precision.fixed
index 8a8c2e1939c8..698238f34365 100644
--- a/tests/ui/excessive_precision.fixed
+++ b/tests/ui/excessive_precision.fixed
@@ -101,4 +101,12 @@ fn main() {
     const _: f64 = 3.0;
     //~^ excessive_precision
     const _: f64 = 3.0000000000000000;
+
+    // Overly specified constants
+    let _: f32 = 1.012_345_7;
+    //~^ excessive_precision
+    let _: f64 = 1.012_345_678_901_234_6;
+    //~^ excessive_precision
+    const _: f32 = 1.01234567890123456789012345678901234567890;
+    const _: f64 = 1.01234567890123456789012345678901234567890;
 }
diff --git a/tests/ui/excessive_precision.rs b/tests/ui/excessive_precision.rs
index 5dcf55cb9273..c8f4f79c1042 100644
--- a/tests/ui/excessive_precision.rs
+++ b/tests/ui/excessive_precision.rs
@@ -101,4 +101,12 @@ fn main() {
     const _: f64 = 3.0000000000000000e+00;
     //~^ excessive_precision
     const _: f64 = 3.0000000000000000;
+
+    // Overly specified constants
+    let _: f32 = 1.01234567890123456789012345678901234567890;
+    //~^ excessive_precision
+    let _: f64 = 1.01234567890123456789012345678901234567890;
+    //~^ excessive_precision
+    const _: f32 = 1.01234567890123456789012345678901234567890;
+    const _: f64 = 1.01234567890123456789012345678901234567890;
 }
diff --git a/tests/ui/excessive_precision.stderr b/tests/ui/excessive_precision.stderr
index f5eeadf0c8cb..e4ddbd12f05e 100644
--- a/tests/ui/excessive_precision.stderr
+++ b/tests/ui/excessive_precision.stderr
@@ -192,5 +192,39 @@ LL -     const _: f64 = 3.0000000000000000e+00;
 LL +     const _: f64 = 3.0;
    |
 
-error: aborting due to 16 previous errors
+error: float has excessive precision
+  --> tests/ui/excessive_precision.rs:106:18
+   |
+LL |     let _: f32 = 1.01234567890123456789012345678901234567890;
+   |                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+   |
+note: consider making it a `const` item
+  --> tests/ui/excessive_precision.rs:106:5
+   |
+LL |     let _: f32 = 1.01234567890123456789012345678901234567890;
+   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+help: consider changing the type or truncating it to
+   |
+LL -     let _: f32 = 1.01234567890123456789012345678901234567890;
+LL +     let _: f32 = 1.012_345_7;
+   |
+
+error: float has excessive precision
+  --> tests/ui/excessive_precision.rs:108:18
+   |
+LL |     let _: f64 = 1.01234567890123456789012345678901234567890;
+   |                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+   |
+note: consider making it a `const` item
+  --> tests/ui/excessive_precision.rs:108:5
+   |
+LL |     let _: f64 = 1.01234567890123456789012345678901234567890;
+   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+help: consider changing the type or truncating it to
+   |
+LL -     let _: f64 = 1.01234567890123456789012345678901234567890;
+LL +     let _: f64 = 1.012_345_678_901_234_6;
+   |
+
+error: aborting due to 18 previous errors