diff --git a/libs/vertexai/langchain_google_vertexai/_base.py b/libs/vertexai/langchain_google_vertexai/_base.py
index 32d895a4..e7ffc646 100644
--- a/libs/vertexai/langchain_google_vertexai/_base.py
+++ b/libs/vertexai/langchain_google_vertexai/_base.py
@@ -105,8 +105,10 @@ class _VertexAIBase(BaseModel):
     def validate_params_base(cls, values: dict) -> Any:
         if "model" in values and "model_name" not in values:
             values["model_name"] = values.pop("model")
-        if values.get("api_transport") is None:
+        if "api_transport" not in values:
             values["api_transport"] = initializer.global_config._api_transport
+        if "location" not in values:
+            values["location"] = initializer.global_config.location
         if values.get("api_endpoint"):
             api_endpoint = values["api_endpoint"]
         else:
diff --git a/libs/vertexai/tests/integration_tests/test_chat_models.py b/libs/vertexai/tests/integration_tests/test_chat_models.py
index b943f6b1..a3eb26db 100644
--- a/libs/vertexai/tests/integration_tests/test_chat_models.py
+++ b/libs/vertexai/tests/integration_tests/test_chat_models.py
@@ -8,6 +8,7 @@
 
 import pytest
 import requests
+import vertexai  # type: ignore[import-untyped, unused-ignore]
 from google.cloud import storage
 from google.cloud.aiplatform_v1beta1.types import Blob, Content, Part
 from google.oauth2 import service_account
@@ -772,6 +773,8 @@ def test_prediction_client_transport():
     assert model.prediction_client.transport.kind == "rest"
     assert model.async_prediction_client.transport.kind == "rest"
 
+    vertexai.init(api_transport="grpc")  # Reset global config to "grpc"
+
 
 @pytest.mark.extended
 def test_structured_output_schema_json():
@@ -1201,3 +1204,23 @@ def test_logprobs() -> None:
     llm3 = ChatVertexAI(model="gemini-1.5-flash", logprobs=False)
     msg3 = llm3.invoke("howdy")
     assert msg3.response_metadata.get("logprobs_result") is None
+
+
+def test_location_init() -> None:
+    # If I don't initialize vertexai before, defaults to us-central-1
+    llm = ChatVertexAI(model="gemini-1.5-flash", logprobs=2)
+    assert llm.location == "us-central1"
+
+    # If I init vertexai with other region the model is in that particular region
+    vertexai.init(location="europe-west1")
+    llm = ChatVertexAI(model="gemini-1.5-flash", logprobs=2)
+    assert llm.location == "europe-west1"
+
+    # If I specify the location, it follows that location
+    llm = ChatVertexAI(model="gemini-1.5-flash", logprobs=2, location="europe-west2")
+    assert llm.location == "europe-west2"
+
+    # It reverts to the default
+    vertexai.init(location="us-central1")
+    llm = ChatVertexAI(model="gemini-1.5-flash", logprobs=2)
+    assert llm.location == "us-central1"
diff --git a/libs/vertexai/tests/integration_tests/test_standard.py b/libs/vertexai/tests/integration_tests/test_standard.py
index f811c57e..c0f52461 100644
--- a/libs/vertexai/tests/integration_tests/test_standard.py
+++ b/libs/vertexai/tests/integration_tests/test_standard.py
@@ -25,6 +25,7 @@ def chat_model_params(self) -> dict:
             "model_name": "gemini-2.0-flash-exp",
             "rate_limiter": rate_limiter,
             "temperature": 0,
+            "api_transport": None,
         }
 
     @property
@@ -52,6 +53,7 @@ def chat_model_params(self) -> dict:
             "model_name": "gemini-1.0-pro-001",
             "rate_limiter": rate_limiter,
             "temperature": 0,
+            "api_transport": None,
         }
 
     @pytest.mark.xfail(reason="Gemini 1.0 doesn't support tool_choice='any'")
@@ -72,6 +74,7 @@ def chat_model_params(self) -> dict:
             "model_name": "gemini-1.5-pro-001",
             "rate_limiter": rate_limiter,
             "temperature": 0,
+            "api_transport": None,
         }
 
     @property