diff --git a/src/rtichoke/_viz_spec_v2.py b/src/rtichoke/_viz_spec_v2.py
index 70992627..8aacf154 100644
--- a/src/rtichoke/_viz_spec_v2.py
+++ b/src/rtichoke/_viz_spec_v2.py
@@ -102,9 +102,12 @@ def _lift_v2_spec_from_performance_data(
     references: list[dict[str, object]] = [
         {"type": "horizontal", "value": 1.0, "scope": "global", "label": "Random"}
     ]
+    perfect_heights: list[float] = []
     for population in populations:
         p = prevalence[population]
         if p > 0:
+            perfect_height = 1.0 / p
+            perfect_heights.append(perfect_height)
             references.append(
                 {
                     "type": "path",
@@ -112,12 +115,16 @@ def _lift_v2_spec_from_performance_data(
                     "population": population,
                     "label": "Perfect Model",
                     "points": [
-                        {"x": 0.0, "y": 1.0 / p},
-                        {"x": p, "y": 1.0 / p},
+                        {"x": 0.0, "y": perfect_height},
+                        {"x": p, "y": perfect_height},
                         {"x": 1.0, "y": 1.0},
                     ],
                 }
             )
+    spec["yAxis"] = {
+        "label": "Lift",
+        "domain": [0, _lift_y_axis_upper_bound(performance_data, perfect_heights)],
+    }
     spec["references"] = references
     return spec
 
@@ -356,8 +363,11 @@ def _lift_times_v2_spec_from_performance_data(
     references: list[dict[str, object]] = [
         {"type": "horizontal", "value": 1.0, "scope": "global", "label": "Random"}
     ]
+    perfect_heights: list[float] = []
     for (population, horizon), risk in risks.items():
         if risk > 0:
+            perfect_height = 1.0 / risk
+            perfect_heights.append(perfect_height)
             references.append(
                 {
                     "type": "path",
@@ -366,8 +376,8 @@ def _lift_times_v2_spec_from_performance_data(
                     "horizon": horizon,
                     "label": "Perfect Model",
                     "points": [
-                        {"x": 0.0, "y": 1.0 / risk},
-                        {"x": risk, "y": 1.0 / risk},
+                        {"x": 0.0, "y": perfect_height},
+                        {"x": risk, "y": perfect_height},
                         {"x": 1.0, "y": 1.0},
                     ],
                 }
@@ -382,7 +392,10 @@ def _lift_times_v2_spec_from_performance_data(
         "x": "ppcr",
         "y": "lift",
         "xAxis": {"label": "Predicted Positives (Rate)", "domain": [0, 1]},
-        "yAxis": {"label": "Lift", "domain": [0, None]},
+        "yAxis": {
+            "label": "Lift",
+            "domain": [0, _lift_y_axis_upper_bound(performance_data, perfect_heights)],
+        },
         "references": references,
     }
 
@@ -473,6 +486,16 @@ def _gains_population_prevalence(
     return prevalence
 
 
+def _lift_y_axis_upper_bound(
+    performance_data: pl.DataFrame,
+    perfect_heights: list[float],
+) -> float:
+    """Return the finite numeric Lift bound implied by existing plot quantities."""
+    observed_lift = performance_data["lift"].max()
+    assert isinstance(observed_lift, (int, float))
+    return max(1.0, float(observed_lift), *perfect_heights)
+
+
 def _curve_v2_spec_from_performance_data(
     performance_data: pl.DataFrame,
     evaluation_metadata: Mapping[str, _EvaluationMetadata],
@@ -609,6 +632,9 @@ def _curve_v2_spec_from_performance_data(
         "x": "ppcr",
         "y": "lift",
         "xAxis": {"label": "Predicted Positives (Rate)", "domain": [0, 1]},
-        "yAxis": {"label": "Lift", "domain": [0, None]},
+        "yAxis": {
+            "label": "Lift",
+            "domain": [0, _lift_y_axis_upper_bound(performance_data, [])],
+        },
         "references": [],
     }
diff --git a/tests/test_lift_v2.py b/tests/test_lift_v2.py
index e41c2bc5..7e242bed 100644
--- a/tests/test_lift_v2.py
+++ b/tests/test_lift_v2.py
@@ -61,6 +61,31 @@ def test_static_lift_v2_spec_single_model():
     ]
 
 
+def test_static_lift_v2_y_axis_domain_is_numeric_and_covers_plot():
+    probs = {
+        "Population A": np.array([0.1, 0.2, 0.7, 0.9]),
+        "Population B": np.array([0.1, 0.3, 0.4, 0.8]),
+    }
+    reals = {
+        "Population A": np.array([0, 0, 1, 1]),
+        "Population B": np.array([0, 0, 0, 1]),
+    }
+    spec = _spec(probs, reals)
+
+    upper = spec["yAxis"]["domain"][1]
+    observed_upper = max(float(row["lift"]) for row in spec["data"])
+    perfect_upper = max(
+        float(point["y"])
+        for reference in spec["references"]
+        if reference["type"] == "path"
+        for point in reference["points"]
+    )
+
+    assert isinstance(upper, (int, float))
+    assert np.isfinite(upper)
+    assert spec["yAxis"]["domain"] == [0, max(1.0, observed_upper, perfect_upper)]
+
+
 def test_static_lift_v2_spec_shared_population():
     probs, reals = _shared_model_inputs()
     spec = _spec(probs, reals)
diff --git a/tests/test_time_lift_v2.py b/tests/test_time_lift_v2.py
index cd31fd60..c7335390 100644
--- a/tests/test_time_lift_v2.py
+++ b/tests/test_time_lift_v2.py
@@ -86,6 +86,24 @@ def test_time_lift_uses_one_evaluation_and_series_per_model_horizon():
     }
 
 
+def test_time_lift_v2_y_axis_domain_is_numeric_and_covers_plot():
+    probs, reals, times = _shared_inputs()
+    spec = _spec(probs, reals, times)
+
+    upper = spec["yAxis"]["domain"][1]
+    observed_upper = max(float(row["lift"]) for row in spec["data"])
+    perfect_upper = max(
+        float(point["y"])
+        for reference in spec["references"]
+        if reference["type"] == "path"
+        for point in reference["points"]
+    )
+
+    assert isinstance(upper, (int, float))
+    assert np.isfinite(upper)
+    assert spec["yAxis"]["domain"] == [0, max(1.0, observed_upper, perfect_upper)]
+
+
 def test_equal_risk_population_horizons_remain_distinct_lift_reference_owners():
     probs = {
         "Population A": np.array([0.05, 0.2, 0.7, 0.95]),