diff --git a/litellm/proxy/management_endpoints/roi_calculator_endpoints.py b/litellm/proxy/management_endpoints/roi_calculator_endpoints.py
index 593d14bfaac..d8a1b781c54 100644
--- a/litellm/proxy/management_endpoints/roi_calculator_endpoints.py
+++ b/litellm/proxy/management_endpoints/roi_calculator_endpoints.py
@@ -6,6 +6,7 @@ from types import MappingProxyType
 from typing import Annotated, Final, Literal
 
 import httpx
+from apscheduler.schedulers.asyncio import AsyncIOScheduler  # pyright: ignore[reportMissingTypeStubs]  # no upstream stubs
 from fastapi import APIRouter, Depends, FastAPI, HTTPException, Query
 from pydantic import BaseModel, ConfigDict, Field, SecretStr, TypeAdapter, ValidationError
 
@@ -556,6 +557,17 @@ def _next_update(settings: ROISettings, status: ROISyncStatus, report: ROIReport
     return utc_anchor + timedelta(minutes=settings.update_interval_minutes)
 
 
+def register_scheduled_sync(scheduler: AsyncIOScheduler) -> None:
+    scheduler.add_job(  # pyright: ignore[reportUnknownMemberType]  # APScheduler exposes untyped scheduling parameters
+        run_scheduled_sync,
+        "interval",
+        seconds=30,
+        id="roi_calculator_refresh",
+        max_instances=1,
+        replace_existing=True,
+    )
+
+
 async def run_scheduled_sync() -> None:
     from litellm.proxy.proxy_server import prisma_client
 
diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py
index 915abf70b6f..83b439188fb 100644
--- a/litellm/proxy/proxy_server.py
+++ b/litellm/proxy/proxy_server.py
@@ -1549,9 +1549,9 @@ async def proxy_startup_event(app: FastAPI) -> AsyncGenerator[None, None]:
             model_info_scheduler.start()
 
     if scheduler is not None and prisma_client is not None:
-        from litellm.proxy.management_endpoints.roi_calculator_endpoints import run_scheduled_sync
+        from litellm.proxy.management_endpoints.roi_calculator_endpoints import register_scheduled_sync
 
-        scheduler.add_job(run_scheduled_sync, "interval", seconds=30, id="roi_calculator_refresh", max_instances=1)
+        register_scheduled_sync(scheduler)
 
     # End of startup event
     yield
diff --git a/tests/unit/proxy/management_endpoints/test_roi_calculator_endpoints.py b/tests/unit/proxy/management_endpoints/test_roi_calculator_endpoints.py
index eaf896831b1..df3a6256965 100644
--- a/tests/unit/proxy/management_endpoints/test_roi_calculator_endpoints.py
+++ b/tests/unit/proxy/management_endpoints/test_roi_calculator_endpoints.py
@@ -6,6 +6,7 @@ from types import MappingProxyType
 from typing import Final, cast
 
 import pytest
+from apscheduler.schedulers.asyncio import AsyncIOScheduler
 from fastapi import FastAPI
 from fastapi.testclient import TestClient
 from pydantic import TypeAdapter
@@ -16,7 +17,9 @@ from litellm.proxy.management_endpoints.roi_calculator_endpoints import (
     _estimator_models_from_deployments,
     _next_update,
     get_roi_config_repository,
+    register_scheduled_sync,
     router,
+    run_scheduled_sync,
 )
 from litellm.proxy.roi_calculator.estimator import estimator_options
 from litellm.proxy.roi_calculator.sample import sample_report
@@ -25,6 +28,21 @@ from litellm.types.roi_calculator import ROIReport, ROISettings, ROISyncStatus
 _JSON_HEADERS: Final = MappingProxyType({"content-type": "application/json"})
 
 
+@pytest.mark.asyncio
+async def test_repeated_startup_keeps_one_roi_schedule() -> None:
+    scheduler: Final = AsyncIOScheduler()
+    scheduler.start(paused=True)
+    try:
+        register_scheduled_sync(scheduler)
+        register_scheduled_sync(scheduler)
+
+        jobs: Final = scheduler.get_jobs()
+        assert len(jobs) == 1
+        assert jobs[0].func is run_scheduled_sync
+    finally:
+        scheduler.shutdown(wait=False)
+
+
 def _assert_json_round_trip(value: object) -> None:
     serialized: Final = json.dumps(value)
     decoded: Final[object] = cast(object, json.loads(serialized))