diff --git a/headroom/proxy/server.py b/headroom/proxy/server.py
index b069335e0..59fe81cfb 100644
--- a/headroom/proxy/server.py
+++ b/headroom/proxy/server.py
@@ -1708,14 +1708,16 @@ class HeadroomProxy(
                 self.warmup.merge_transform_status(transform_status)
 
         # Update internal status from eager loading results
-        if eager_status.get("kompress") == "enabled":
-            self._kompress_status = "enabled"
+        if eager_status.get("kompress") in {"enabled", "deferred"}:
+            self._kompress_status = eager_status["kompress"]
         if eager_status.get("code_aware") == "enabled":
             self._code_aware_status = "enabled"
 
         # Log component status
         if self._kompress_status == "enabled":
             logger.info("Kompress: ENABLED (ModernBERT token compressor)")
+        elif self._kompress_status == "deferred":
+            logger.info("Kompress: DEFERRED (model loads on first request)")
         elif self.config.optimize:
             logger.info("Kompress: not installed (pip install headroom-ai[ml] for ML compression)")
 
@@ -2630,6 +2632,21 @@ def create_app(config: ProxyConfig | None = None) -> FastAPI:
                 return True
             proxy.warmup.kompress.mark_loaded(handle=compressor, backend=backend)
             return True
+
+        try:
+            from headroom.transforms.kompress_compressor import HF_MODEL_ID, _kompress_cache
+        except ImportError:
+            return True
+
+        cached = _kompress_cache.get(HF_MODEL_ID)
+        if cached is not None:
+            try:
+                model, _tokenizer, backend = cached
+            except (TypeError, ValueError):
+                pass
+            else:
+                if backend and proxy.warmup.kompress.status != "loaded":
+                    proxy.warmup.kompress.mark_loaded(handle=model, backend=backend)
         return True
 
     def _health_checks() -> dict[str, dict[str, Any]]:
diff --git a/tests/test_proxy_eager_preload_bind.py b/tests/test_proxy_eager_preload_bind.py
index c168197a4..45b05c549 100644
--- a/tests/test_proxy_eager_preload_bind.py
+++ b/tests/test_proxy_eager_preload_bind.py
@@ -11,6 +11,7 @@ bind still happens and transforms fall back to lazy loading.
 
 from __future__ import annotations
 
+import logging
 import threading
 import time
 
@@ -118,3 +119,25 @@ async def test_startup_merges_warmup_for_normal_transforms(monkeypatch):
         assert proxy._kompress_status == "enabled"
     finally:
         await proxy.shutdown()
+
+
+async def test_startup_reports_deferred_kompress(caplog):
+    proxy = _make_proxy(optimize=True)
+    proxy.anthropic_pipeline = _FakePipeline([_FastTransform({"kompress": "deferred"})])
+    proxy.openai_pipeline = _FakePipeline([])
+
+    try:
+        # Proxy setup disables propagation on the ``headroom`` logger, so
+        # attach caplog's handler directly to the logger that emits this line.
+        server_mod.logger.addHandler(caplog.handler)
+        try:
+            with caplog.at_level(logging.INFO, logger=server_mod.logger.name):
+                await proxy.startup()
+        finally:
+            server_mod.logger.removeHandler(caplog.handler)
+
+        assert proxy._kompress_status == "deferred"
+        assert "Kompress: DEFERRED (model loads on first request)" in caplog.messages
+        assert not any("Kompress: not installed" in message for message in caplog.messages)
+    finally:
+        await proxy.shutdown()
diff --git a/tests/test_proxy_health.py b/tests/test_proxy_health.py
index af459535e..e2a2d89d1 100644
--- a/tests/test_proxy_health.py
+++ b/tests/test_proxy_health.py
@@ -3,6 +3,7 @@ from fastapi.testclient import TestClient
 
 from headroom.proxy.models import ProxyConfig
 from headroom.proxy.server import create_app
+from headroom.transforms import kompress_compressor
 
 
 class _ReadyCompressor:
@@ -88,6 +89,31 @@ def test_readyz_promotes_deferred_kompress_after_runtime_load(monkeypatch):
     }
 
 
+@pytest.mark.parametrize("attached", [False, True])
+def test_readyz_promotes_kompress_from_module_cache(monkeypatch, attached):
+    model = object()
+    monkeypatch.setattr(
+        kompress_compressor,
+        "_kompress_cache",
+        {kompress_compressor.HF_MODEL_ID: (model, object(), "onnx")},
+    )
+    compressor = _ReadyCompressor(ready=False) if attached else None
+    app, proxy = _health_app(monkeypatch, compressor)
+    proxy.warmup.kompress.info["source_status"] = "deferred"
+
+    payload = TestClient(app).get("/readyz").json()
+
+    assert payload["checks"]["kompress"] == {
+        "enabled": True,
+        "ready": True,
+        "status": "healthy",
+        "backend": "onnx",
+    }
+    assert proxy.warmup.kompress.handle is model
+    if compressor is not None:
+        assert compressor.calls == ["is_ready"]
+
+
 def test_readyz_promotes_remote_kompress_backend(monkeypatch):
     compressor = _ReadyCompressor(backend="remote")
     app, proxy = _health_app(monkeypatch)