diff --git a/.circleci/run_xpu_tests.py b/.circleci/run_xpu_tests.py
index ef91b82540..364b80ecea 100644
--- a/.circleci/run_xpu_tests.py
+++ b/.circleci/run_xpu_tests.py
@@ -32,7 +32,7 @@
         '-d', '--device',
         type=str,
         help='device name',
-        choices=['host', 'cpu', 'gpu']
+        choices=['cpu', 'gpu']
     )
     parser.add_argument(
         '--deselect',
diff --git a/daal4py/sklearn/_utils.py b/daal4py/sklearn/_utils.py
index 47aff5d8d2..6528805ecd 100644
--- a/daal4py/sklearn/_utils.py
+++ b/daal4py/sklearn/_utils.py
@@ -115,13 +115,13 @@ def get_patch_message(s):
         if 'daal4py.oneapi' in sys.modules:
             from daal4py.oneapi import _get_device_name_sycl_ctxt
             dev = _get_device_name_sycl_ctxt()
-            if dev == 'cpu' or dev == 'host' or dev is None:
+            if dev == 'cpu' or dev is None:
                 message += 'CPU'
             elif dev == 'gpu':
                 message += 'GPU'
             else:
                 raise ValueError(f"Unexpected device name {dev}."
-                                 " Supported types are host, cpu and gpu")
+                                 " Supported types are cpu and gpu")
         else:
             message += 'CPU'
 
diff --git a/examples/daal4py/run_examples.py b/examples/daal4py/run_examples.py
index ad7bec3066..14eba0a21b 100755
--- a/examples/daal4py/run_examples.py
+++ b/examples/daal4py/run_examples.py
@@ -72,7 +72,6 @@
             availabe_devices.append("gpu")
     except RuntimeError:
         gpu_available = False
-    availabe_devices.append("host")
     availabe_devices.append("cpu")
     #validate that host and cpu devices avaialbe for logging reasons. Examples and
     #vaidaton logic assumes that host and cpu devices are always available
diff --git a/examples/daal4py/sycl/sklearn_sycl.py b/examples/daal4py/sycl/sklearn_sycl.py
index 933accc765..c33ded4751 100644
--- a/examples/daal4py/sycl/sklearn_sycl.py
+++ b/examples/daal4py/sycl/sklearn_sycl.py
@@ -141,15 +141,13 @@ def get_context(device):
 
 def device_type_to_str(queue):
     if queue is None:
-        return 'host'
+        return 'cpu'
 
     from dpctl import device_type
     if queue.sycl_device.device_type == device_type.cpu:
         return 'cpu'
     if queue.sycl_device.device_type == device_type.gpu:
         return 'gpu'
-    if queue.sycl_device.device_type == device_type.host:
-        return 'host'
     return 'unknown'
 
 
diff --git a/onedal/common/_policy.py b/onedal/common/_policy.py
index 5d6c51d011..8350874184 100644
--- a/onedal/common/_policy.py
+++ b/onedal/common/_policy.py
@@ -46,7 +46,7 @@ def __init__(self):
         if 'daal4py.oneapi' in sys.modules:
             from daal4py.oneapi import _get_sycl_ctxt, sycl_execution_context
             self._d4p_context = _get_sycl_ctxt()
-            self._host_context = sycl_execution_context('host')
+            self._host_context = sycl_execution_context('cpu')
             self._host_context.apply()
 
     def __del__(self):
diff --git a/onedal/common/policy.cpp b/onedal/common/policy.cpp
index fd1a925d34..42a117400b 100644
--- a/onedal/common/policy.cpp
+++ b/onedal/common/policy.cpp
@@ -25,7 +25,7 @@ ONEDAL_PY_INIT_MODULE(policy) {
     py::class_<detail::host_policy>(m, "host_policy")
         .def(py::init())
         .def("get_device_name", [](const detail::host_policy& self) {
-            return "host";
+            return "cpu";
         });
 
 #ifdef ONEDAL_DATA_PARALLEL
@@ -43,8 +43,6 @@ ONEDAL_PY_INIT_MODULE(policy) {
                 return "gpu";
             } else if (self.get_queue().get_device().is_cpu()) {
                 return "cpu";
-            } else if (self.get_queue().get_device().is_host()) {
-                return "host";
             }
             return "unknown";
         });
diff --git a/onedal/datatypes/data_conversion.cpp b/onedal/datatypes/data_conversion.cpp
index ebfcd28a0b..00f406afb5 100644
--- a/onedal/datatypes/data_conversion.cpp
+++ b/onedal/datatypes/data_conversion.cpp
@@ -34,7 +34,7 @@ static dal::array<T> transfer_to_host(const dal::array<T>& array) {
     auto opt_queue = array.get_queue();
     if (opt_queue.has_value()) {
         auto device = opt_queue->get_device();
-        if (!device.is_cpu() && !device.is_host()) {
+        if (!device.is_cpu()) {
             const auto* device_data = array.get_data();
 
             auto memory_kind = sycl::get_pointer_type(device_data, opt_queue->get_context());
diff --git a/sklearnex/_device_offload.py b/sklearnex/_device_offload.py
index 281854773f..f580acbb7c 100644
--- a/sklearnex/_device_offload.py
+++ b/sklearnex/_device_offload.py
@@ -37,9 +37,8 @@ def __init__(self, filter_string):
             self._filter_string = filter_string
             self.is_cpu = 'cpu' in filter_string
             self.is_gpu = 'gpu' in filter_string
-            self.is_host = False
 
-            if not (self.is_cpu or self.is_host):
+            if not (self.is_cpu):
                 import logging
                 logging.warning("Device support is limited. "
                                 "Please install dpctl for full experience")
diff --git a/sklearnex/_utils.py b/sklearnex/_utils.py
index 1cb569c3aa..a03bf774d1 100755
--- a/sklearnex/_utils.py
+++ b/sklearnex/_utils.py
@@ -47,7 +47,7 @@ def get_patch_message(s, queue=None, cpu_fallback=False):
         elif 'daal4py.oneapi' in sys.modules:
             from daal4py.oneapi import _get_device_name_sycl_ctxt
             dev = _get_device_name_sycl_ctxt()
-            if dev == 'cpu' or dev == 'host' or dev is None:
+            if dev == 'cpu' or dev is None:
                 message += 'CPU'
             elif dev == 'gpu':
                 if cpu_fallback: