diff --git a/doc/whats_new/v1.1.rst b/doc/whats_new/v1.1.rst
index cd1102b77ed2e..471c6b86eec2d 100644
--- a/doc/whats_new/v1.1.rst
+++ b/doc/whats_new/v1.1.rst
@@ -258,6 +258,10 @@ Changelog
   Setting a transformer to "passthrough" will pass the features unchanged.
   :pr:`20860` by :user:`Shubhraneel Pal <shubhraneel>`.
 
+- |Fix| :class: `pipeline.Pipeline` now does not validate hyper-parameters in
+  `__init__` but in `.fit()`.
+  :pr:`21888` by :user:`iofall <iofall>` and :user: `Arisa Y. <arisayosh>`.
+
 :mod:`sklearn.preprocessing`
 ............................
 
diff --git a/sklearn/pipeline.py b/sklearn/pipeline.py
index 6134b6318c838..794919f6942e1 100644
--- a/sklearn/pipeline.py
+++ b/sklearn/pipeline.py
@@ -146,7 +146,6 @@ def __init__(self, steps, *, memory=None, verbose=False):
         self.steps = steps
         self.memory = memory
         self.verbose = verbose
-        self._validate_steps()
 
     def get_params(self, deep=True):
         """Get parameters for this estimator.
diff --git a/sklearn/tests/test_common.py b/sklearn/tests/test_common.py
index 8b0796b2b5dbc..80b9a71f853fc 100644
--- a/sklearn/tests/test_common.py
+++ b/sklearn/tests/test_common.py
@@ -428,7 +428,6 @@ def test_transformers_get_feature_names_out(transformer):
     "FeatureUnion",
     "GridSearchCV",
     "HalvingGridSearchCV",
-    "Pipeline",
     "SGDOneClassSVM",
     "TheilSenRegressor",
     "TweedieRegressor",
diff --git a/sklearn/tests/test_pipeline.py b/sklearn/tests/test_pipeline.py
index 77201628957e0..3a94bf10b6d3c 100644
--- a/sklearn/tests/test_pipeline.py
+++ b/sklearn/tests/test_pipeline.py
@@ -165,20 +165,23 @@ def predict_log_proba(self, X, got_attribute=False):
         return self
 
 
-def test_pipeline_init():
-    # Test the various init parameters of the pipeline.
+def test_pipeline_invalid_parameters():
+    # Test the various init parameters of the pipeline in fit
+    # method
+    pipeline = Pipeline([(1, 1)])
     with pytest.raises(TypeError):
-        Pipeline()
+        pipeline.fit([[1]], [1])
 
-    # Check that we can't instantiate pipelines with objects without fit
+    # Check that we can't fit pipelines with objects without fit
     # method
     msg = (
         "Last step of Pipeline should implement fit "
         "or be the string 'passthrough'"
         ".*NoFit.*"
     )
+    pipeline = Pipeline([("clf", NoFit())])
     with pytest.raises(TypeError, match=msg):
-        Pipeline([("clf", NoFit())])
+        pipeline.fit([[1]], [1])
 
     # Smoke test with only an estimator
     clf = NoTrans()
@@ -203,11 +206,12 @@ def test_pipeline_init():
     assert pipe.named_steps["anova"] is filter1
     assert pipe.named_steps["svc"] is clf
 
-    # Check that we can't instantiate with non-transformers on the way
+    # Check that we can't fit with non-transformers on the way
     # Note that NoTrans implements fit, but not transform
     msg = "All intermediate steps should be transformers.*\\bNoTrans\\b.*"
+    pipeline = Pipeline([("t", NoTrans()), ("svc", clf)])
     with pytest.raises(TypeError, match=msg):
-        Pipeline([("t", NoTrans()), ("svc", clf)])
+        pipeline.fit([[1]], [1])
 
     # Check that params are set
     pipe.set_params(svc__C=0.1)
@@ -1086,7 +1090,7 @@ def test_step_name_validation():
             # three ways to make invalid:
             # - construction
             with pytest.raises(ValueError, match=message):
-                cls(**{param: bad_steps})
+                cls(**{param: bad_steps}).fit([[1]], [1])
 
             # - setattr
             est = cls(**{param: [("a", Mult(1))]})