diff --git a/src/tests/end2end/ShaderFloat16Tests.cpp b/src/tests/end2end/ShaderFloat16Tests.cpp
index 1f4f445ef5..c920f99b30 100644
--- a/src/tests/end2end/ShaderFloat16Tests.cpp
+++ b/src/tests/end2end/ShaderFloat16Tests.cpp
@@ -38,6 +38,7 @@ class ShaderFloat16Tests : public DawnTest {
 // Test basic 16bit float arithmetic and 16bit storage features.
 TEST_P(ShaderFloat16Tests, Basic16BitFloatFeaturesTest) {
     DAWN_SKIP_TEST_IF(!IsShaderFloat16Supported());
+    DAWN_SKIP_TEST_IF(IsD3D12() && IsIntel());  // Flaky crashes. crbug.com/dawn/586
 
     uint16_t uniformData[] = {Float32ToFloat16(1.23), Float32ToFloat16(0.0)};  // 0.0 is a padding.
     wgpu::Buffer uniformBuffer = utils::CreateBufferFromData(