diff --git a/crates/polars-core/src/fmt.rs b/crates/polars-core/src/fmt.rs
index 2ba0cee9f4f47..d47f4722b88f7 100644
--- a/crates/polars-core/src/fmt.rs
+++ b/crates/polars-core/src/fmt.rs
@@ -2,6 +2,7 @@
 use std::borrow::Cow;
 use std::fmt::{Debug, Display, Formatter};
 use std::sync::atomic::{AtomicU8, Ordering};
+use std::sync::RwLock;
 use std::{fmt, str};
 
 #[cfg(any(
@@ -33,7 +34,7 @@ pub enum FloatFmt {
     Full,
 }
 static FLOAT_FMT: AtomicU8 = AtomicU8::new(FloatFmt::Mixed as u8);
-static FLOAT_PRECISION: AtomicU8 = AtomicU8::new(u8::MAX);
+static FLOAT_PRECISION: RwLock<Option<usize>> = RwLock::new(None);
 
 pub fn get_float_fmt() -> FloatFmt {
     match FLOAT_FMT.load(Ordering::Relaxed) {
@@ -43,16 +44,16 @@ pub fn get_float_fmt() -> FloatFmt {
     }
 }
 
-pub fn get_float_precision() -> u8 {
-    FLOAT_PRECISION.load(Ordering::Relaxed)
+pub fn get_float_precision() -> Option<usize> {
+    FLOAT_PRECISION.read().unwrap().clone()
 }
 
 pub fn set_float_fmt(fmt: FloatFmt) {
     FLOAT_FMT.store(fmt as u8, Ordering::Relaxed)
 }
 
-pub fn set_float_precision(precision: u8) {
-    FLOAT_PRECISION.store(precision, Ordering::Relaxed)
+pub fn set_float_precision(precision: Option<usize>) {
+    *FLOAT_PRECISION.write().unwrap() = precision;
 }
 
 macro_rules! format_array {
@@ -724,12 +725,16 @@ const SCIENTIFIC_BOUND: f64 = 999999.0;
 fn fmt_float<T: Num + NumCast>(f: &mut Formatter<'_>, width: usize, v: T) -> fmt::Result {
     let v: f64 = NumCast::from(v).unwrap();
 
-    let precision = get_float_precision();
-    if precision != u8::MAX {
-        if format!("{v:.precision$}", precision = precision as usize).len() > 19 {
-            return write!(f, "{v:>width$.precision$e}", precision = precision as usize);
-        }
-        return write!(f, "{v:>width$.precision$}", precision = precision as usize);
+    let float_precision = get_float_precision();
+
+    match float_precision {
+        Some(precision) => {
+            if format!("{v:.precision$}", precision = precision).len() > 19 {
+                return write!(f, "{v:>width$.precision$e}", precision = precision);
+            }
+            return write!(f, "{v:>width$.precision$}", precision = precision);
+        },
+        _ => {},
     }
 
     if matches!(get_float_fmt(), FloatFmt::Full) {
diff --git a/py-polars/polars/config.py b/py-polars/polars/config.py
index 2004b0834d79b..fa5f1a756729c 100644
--- a/py-polars/polars/config.py
+++ b/py-polars/polars/config.py
@@ -17,7 +17,7 @@ def _get_float_fmt() -> str:  # pragma: no cover
 
 
 def _get_float_precision() -> int:
-    return 0
+    return -1
 
 
 # note: module not available when building docs
@@ -379,12 +379,6 @@ def set_float_precision(cls, precision: int | None = None) -> type[Config]:
             Number of decimal places to display
 
         """
-        if precision is None:
-            precision = 255
-        elif precision > 16 and precision != 255:
-            raise ValueError(
-                f"precision must be None, or a number between 0-16; got {precision}"
-            )
         _set_float_precision(precision)
         return cls
 
diff --git a/py-polars/src/functions/meta.rs b/py-polars/src/functions/meta.rs
index 36aecc06633d1..70e408d37f9d3 100644
--- a/py-polars/src/functions/meta.rs
+++ b/py-polars/src/functions/meta.rs
@@ -58,19 +58,14 @@ pub fn get_float_fmt() -> PyResult<String> {
 }
 
 #[pyfunction]
-pub fn set_float_precision(precision: u8) -> PyResult<()> {
+pub fn set_float_precision(precision: Option<usize>) -> PyResult<()> {
     use polars_core::fmt::set_float_precision;
-    if precision > 16 && precision != u8::MAX {
-        return Err(PyValueError::new_err(format!(
-            "maximum supported float precision is 16, got {precision}",
-        )));
-    }
     set_float_precision(precision);
     Ok(())
 }
 
 #[pyfunction]
-pub fn get_float_precision() -> PyResult<u8> {
+pub fn get_float_precision() -> PyResult<Option<usize>> {
     use polars_core::fmt::get_float_precision;
     Ok(get_float_precision())
 }
diff --git a/py-polars/tests/unit/test_cfg.py b/py-polars/tests/unit/test_cfg.py
index f722c129feb80..d926a943adb15 100644
--- a/py-polars/tests/unit/test_cfg.py
+++ b/py-polars/tests/unit/test_cfg.py
@@ -611,10 +611,6 @@ def test_numeric_right_alignment() -> None:
             "└─────────┴─────────┴───────────┘"
         )
 
-    # test nonsensical float precision raises an error
-    with pytest.raises(ValueError):
-        pl.Config.set_float_precision(50)
-
 
 def test_string_cache() -> None:
     df1 = pl.DataFrame({"a": ["foo", "bar", "ham"], "b": [1, 2, 3]})
@@ -684,7 +680,7 @@ def test_config_load_save(tmp_path: Path) -> None:
         assert os.environ.get("POLARS_FMT_MAX_COLS") is None
         assert os.environ.get("POLARS_VERBOSE") is None
         assert _get_float_fmt() == "mixed"
-        assert _get_float_precision() == 255
+        assert _get_float_precision() is None
 
 
 def test_config_scope() -> None: