diff --git a/include/ccmath/internal/math/generic/func/power/impl/hypot_impl.hpp b/include/ccmath/internal/math/generic/func/power/impl/hypot_impl.hpp
index e2cab091..645c9a96 100644
--- a/include/ccmath/internal/math/generic/func/power/impl/hypot_impl.hpp
+++ b/include/ccmath/internal/math/generic/func/power/impl/hypot_impl.hpp
@@ -17,6 +17,7 @@
 #include "ccmath/math/compare/isnan.hpp"
 #include "ccmath/math/power/sqrt.hpp"
 
+#include <limits>
 #include <type_traits>
 
 namespace ccm::internal::impl
@@ -44,6 +45,14 @@ namespace ccm::internal::impl
 		}
 
 		const T ratio = y / x;
-		return x * ccm::sqrt(static_cast<T>(1) + ratio * ratio);
+
+		// Compute scale in double precision to avoid float rounding scale to exactly
+		// 1.0 when ratio is tiny, which would make the overflow guard silently fail.
+		const double ratio_d = static_cast<double>(ratio);
+		const double scale_d = ccm::sqrt(1.0 + ratio_d * ratio_d);
+
+		if (static_cast<double>(x) > static_cast<double>(std::numeric_limits<T>::max()) / scale_d) { return fp_bits_t::inf().get_val(); }
+
+		return x * static_cast<T>(scale_d);
 	}
 } // namespace ccm::internal::impl
diff --git a/tests/src/math/power/hypot_cbrt_test.cpp b/tests/src/math/power/hypot_cbrt_test.cpp
index 9c5b71cf..7e6c468c 100644
--- a/tests/src/math/power/hypot_cbrt_test.cpp
+++ b/tests/src/math/power/hypot_cbrt_test.cpp
@@ -103,3 +103,13 @@ TEST(CcmathPowerTests, HypotCbrtCompileTime)
 	static_assert(ccm::hypot(2.0, 3.0, 6.0) == 7.0);
 	static_assert(ccm::cbrt(8.0) == 2.0);
 }
+TEST(CcmathPowerTests, HypotOverflowBoundary)
+{
+    constexpr float x = 3.40282347e38f;
+    constexpr float y = 9.18e34f;
+
+    ccm::test::ExpectSameFloatingAsStd(
+        ccm::hypot(x, y),
+        std::hypot(x, y)
+    );
+}