diff --git a/src/nvidia-gpu/NvidiaGpuDevice.cpp b/src/nvidia-gpu/NvidiaGpuDevice.cpp
index ab476da..03550e5 100644
--- a/src/nvidia-gpu/NvidiaGpuDevice.cpp
+++ b/src/nvidia-gpu/NvidiaGpuDevice.cpp
@@ -14,6 +14,7 @@
 #include <bits/basic_string.h>
 
 #include <MctpRequester.hpp>
+#include <NvidiaGpuThresholds.hpp>
 #include <boost/asio/io_context.hpp>
 #include <phosphor-logging/lg2.hpp>
 #include <sdbusplus/asio/connection.hpp>
@@ -23,6 +24,7 @@
 #include <cstdint>
 #include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 
 GpuDevice::GpuDevice(const SensorConfigs& configs, const std::string& name,
@@ -45,9 +47,31 @@
         conn, mctpRequester, name + "_TEMP_0", path, eid, gpuTempSensorId,
         objectServer, std::vector<thresholds::Threshold>{});
 
-    tLimitSensor = std::make_shared<NvidiaGpuTempSensor>(
-        conn, mctpRequester, name + "_TEMP_1", path, eid, gpuTLimitSensorId,
-        objectServer, std::vector<thresholds::Threshold>{});
+    readThermalParameters(
+        eid,
+        std::vector<gpuThresholdId>{gpuTLimitWarnringThresholdId,
+                                    gpuTLimitCriticalThresholdId,
+                                    gpuTLimitHardshutDownThresholdId},
+        mctpRequester, [this](uint8_t rc, std::vector<int32_t> thresholds) {
+            std::vector<thresholds::Threshold> tLimitThresholds{};
+            if (rc == 0)
+            {
+                tLimitThresholds = {
+                    thresholds::Threshold{thresholds::Level::WARNING,
+                                          thresholds::Direction::LOW,
+                                          static_cast<double>(thresholds[0])},
+                    thresholds::Threshold{thresholds::Level::CRITICAL,
+                                          thresholds::Direction::LOW,
+                                          static_cast<double>(thresholds[1])},
+                    thresholds::Threshold{thresholds::Level::HARDSHUTDOWN,
+                                          thresholds::Direction::LOW,
+                                          static_cast<double>(thresholds[2])}};
+            }
+
+            tLimitSensor = std::make_shared<NvidiaGpuTempSensor>(
+                conn, mctpRequester, name + "_TEMP_1", path, eid,
+                gpuTLimitSensorId, objectServer, std::move(tLimitThresholds));
+        });
 
     lg2::info("Added GPU {NAME} Sensors with chassis path: {PATH}.", "NAME",
               name, "PATH", path);
@@ -58,7 +82,10 @@
 void GpuDevice::read()
 {
     tempSensor->update();
-    tLimitSensor->update();
+    if (tLimitSensor)
+    {
+        tLimitSensor->update();
+    }
 
     waitTimer.expires_after(std::chrono::milliseconds(sensorPollMs));
     waitTimer.async_wait([this](const boost::system::error_code& ec) {
diff --git a/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp b/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp
index 17f71e0..7a48b30 100644
--- a/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp
+++ b/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp
@@ -152,5 +152,70 @@
 
     return 0;
 }
+
+int encodeReadThermalParametersRequest(uint8_t instanceId, uint8_t sensorId,
+                                       std::span<uint8_t> buf)
+{
+    if (buf.size() < sizeof(ReadThermalParametersRequest))
+    {
+        return EINVAL;
+    }
+
+    auto* msg = reinterpret_cast<ReadThermalParametersRequest*>(buf.data());
+
+    ocp::accelerator_management::BindingPciVidInfo header{};
+    header.ocp_accelerator_management_msg_type =
+        static_cast<uint8_t>(ocp::accelerator_management::MessageType::REQUEST);
+    header.instance_id = instanceId &
+                         ocp::accelerator_management::instanceIdBitMask;
+    header.msg_type = static_cast<uint8_t>(MessageType::PLATFORM_ENVIRONMENTAL);
+
+    auto rc = packHeader(header, msg->hdr.msgHdr.hdr);
+
+    if (rc != 0)
+    {
+        return rc;
+    }
+
+    msg->hdr.command = static_cast<uint8_t>(
+        PlatformEnvironmentalCommands::READ_THERMAL_PARAMETERS);
+    msg->hdr.data_size = sizeof(sensorId);
+    msg->sensor_id = sensorId;
+
+    return 0;
+}
+
+int decodeReadThermalParametersResponse(
+    std::span<const uint8_t> buf,
+    ocp::accelerator_management::CompletionCode& cc, uint16_t& reasonCode,
+    int32_t& threshold)
+{
+    auto rc =
+        ocp::accelerator_management::decodeReasonCodeAndCC(buf, cc, reasonCode);
+
+    if (rc != 0 || cc != ocp::accelerator_management::CompletionCode::SUCCESS)
+    {
+        return rc;
+    }
+
+    if (buf.size() < sizeof(ReadThermalParametersResponse))
+    {
+        return EINVAL;
+    }
+
+    const auto* response =
+        reinterpret_cast<const ReadThermalParametersResponse*>(buf.data());
+
+    uint16_t dataSize = le16toh(response->hdr.data_size);
+
+    if (dataSize != sizeof(int32_t))
+    {
+        return EINVAL;
+    }
+
+    threshold = le32toh(response->threshold);
+
+    return 0;
+}
 // NOLINTEND(cppcoreguidelines-pro-type-reinterpret-cast)
 } // namespace gpu
diff --git a/src/nvidia-gpu/NvidiaGpuMctpVdm.hpp b/src/nvidia-gpu/NvidiaGpuMctpVdm.hpp
index ce3b393..f7c78b8 100644
--- a/src/nvidia-gpu/NvidiaGpuMctpVdm.hpp
+++ b/src/nvidia-gpu/NvidiaGpuMctpVdm.hpp
@@ -30,6 +30,7 @@
 enum class PlatformEnvironmentalCommands : uint8_t
 {
     GET_TEMPERATURE_READING = 0x00,
+    READ_THERMAL_PARAMETERS = 0x02,
 };
 
 enum class DeviceIdentification : uint8_t
@@ -57,12 +58,20 @@
 
 using GetTemperatureReadingRequest = GetNumericSensorReadingRequest;
 
+using ReadThermalParametersRequest = GetNumericSensorReadingRequest;
+
 struct GetTemperatureReadingResponse
 {
     ocp::accelerator_management::CommonResponse hdr;
     int32_t reading;
 } __attribute__((packed));
 
+struct ReadThermalParametersResponse
+{
+    ocp::accelerator_management::CommonResponse hdr;
+    int32_t threshold;
+} __attribute__((packed));
+
 int packHeader(const ocp::accelerator_management::BindingPciVidInfo& hdr,
                ocp::accelerator_management::BindingPciVid& msg);
 
@@ -82,4 +91,12 @@
     ocp::accelerator_management::CompletionCode& cc, uint16_t& reasonCode,
     double& temperatureReading);
 
+int encodeReadThermalParametersRequest(uint8_t instanceId, uint8_t sensorId,
+                                       std::span<uint8_t> buf);
+
+int decodeReadThermalParametersResponse(
+    std::span<const uint8_t> buf,
+    ocp::accelerator_management::CompletionCode& cc, uint16_t& reasonCode,
+    int32_t& threshold);
+
 } // namespace gpu
diff --git a/src/nvidia-gpu/NvidiaGpuThresholds.cpp b/src/nvidia-gpu/NvidiaGpuThresholds.cpp
new file mode 100644
index 0000000..1bb9c0d
--- /dev/null
+++ b/src/nvidia-gpu/NvidiaGpuThresholds.cpp
@@ -0,0 +1,127 @@
+/*
+ * SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION &
+ * AFFILIATES. All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#include "NvidiaGpuThresholds.hpp"
+
+#include <MctpRequester.hpp>
+#include <NvidiaGpuMctpVdm.hpp>
+#include <OcpMctpVdm.hpp>
+#include <phosphor-logging/lg2.hpp>
+
+#include <array>
+#include <cerrno>
+#include <cstddef>
+#include <cstdint>
+#include <functional>
+#include <memory>
+#include <span>
+#include <vector>
+
+void processReadThermalParameterResponse(
+    const std::function<void(uint8_t, int32_t)>& callback,
+    const std::span<const uint8_t> respMsg, int sendRecvMsgResult)
+{
+    if (sendRecvMsgResult != 0)
+    {
+        lg2::error(
+            "Error reading thermal parameter: sending message over MCTP failed, rc={RC}",
+            "RC", sendRecvMsgResult);
+        callback(EPROTO, 0);
+        return;
+    }
+
+    ocp::accelerator_management::CompletionCode cc{};
+    uint16_t reasonCode = 0;
+    int32_t threshold = 0;
+
+    auto rc = gpu::decodeReadThermalParametersResponse(respMsg, cc, reasonCode,
+                                                       threshold);
+
+    if (rc != 0 || cc != ocp::accelerator_management::CompletionCode::SUCCESS)
+    {
+        lg2::error(
+            "Error reading thermal parameter: decode failed, rc={RC}, cc={CC}, reasonCode={RESC}",
+            "RC", rc, "CC", cc, "RESC", reasonCode);
+        callback(EPROTO, 0);
+        return;
+    }
+
+    callback(0, threshold);
+};
+
+void readThermalParameter(uint8_t eid, uint8_t id,
+                          mctp::MctpRequester& mctpRequester,
+                          const std::function<void(uint8_t, int32_t)>& callback)
+{
+    static std::array<uint8_t, sizeof(gpu::ReadThermalParametersRequest)>
+        reqMsg{};
+
+    static std::array<uint8_t, sizeof(gpu::ReadThermalParametersResponse)>
+        respMsg{};
+
+    auto rc = gpu::encodeReadThermalParametersRequest(0, id, reqMsg);
+    if (rc != 0)
+    {
+        lg2::error(
+            "Error reading thermal parameter for eid {EID} and parameter id {PID} : encode failed. rc={RC}",
+            "EID", eid, "PID", id, "RC", rc);
+        callback(rc, 0);
+        return;
+    }
+
+    mctpRequester.sendRecvMsg(
+        eid, reqMsg, respMsg,
+        std::bind_front(processReadThermalParameterResponse, callback,
+                        std::span{respMsg}));
+}
+
+void readThermalParameterCallback(
+    uint8_t eid, const std::shared_ptr<std::vector<uint8_t>>& ids,
+    mctp::MctpRequester& mctpRequester,
+    const std::function<void(uint8_t, std::vector<int32_t>)>& callback,
+    size_t index, const std::shared_ptr<std::vector<int32_t>>& thresholds,
+    uint8_t rc, int32_t threshold)
+{
+    if (rc != 0)
+    {
+        lg2::error(
+            "Error reading thermal parameter for eid {EID} and parameter id {PID}. rc={RC}",
+            "EID", eid, "PID", (*ids)[index], "RC", rc);
+        return;
+        callback(rc, *thresholds);
+        return;
+    }
+
+    thresholds->push_back(threshold);
+
+    ++index;
+    if (index == ids->size())
+    {
+        callback(rc, *thresholds);
+    }
+    else
+    {
+        readThermalParameter(eid, (*ids)[index], mctpRequester,
+                             std::bind_front(readThermalParameterCallback, eid,
+                                             ids, std::ref(mctpRequester),
+                                             callback, index, thresholds));
+    }
+}
+
+void readThermalParameters(
+    uint8_t eid, const std::vector<uint8_t>& ids,
+    mctp::MctpRequester& mctpRequester,
+    const std::function<void(uint8_t, std::vector<int32_t>)>& callback)
+{
+    auto thresholds = std::make_shared<std::vector<int32_t>>();
+    size_t index = 0;
+
+    readThermalParameter(
+        eid, ids[index], mctpRequester,
+        std::bind_front(readThermalParameterCallback, eid,
+                        std::make_shared<std::vector<uint8_t>>(ids),
+                        std::ref(mctpRequester), callback, index, thresholds));
+}
diff --git a/src/nvidia-gpu/NvidiaGpuThresholds.hpp b/src/nvidia-gpu/NvidiaGpuThresholds.hpp
new file mode 100644
index 0000000..9d1970f
--- /dev/null
+++ b/src/nvidia-gpu/NvidiaGpuThresholds.hpp
@@ -0,0 +1,24 @@
+/*
+ * SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION &
+ * AFFILIATES. All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#pragma once
+
+#include "MctpRequester.hpp"
+
+#include <cstdint>
+#include <functional>
+#include <vector>
+
+using gpuThresholdId = uint8_t;
+
+constexpr gpuThresholdId gpuTLimitCriticalThresholdId{1};
+constexpr gpuThresholdId gpuTLimitWarnringThresholdId{2};
+constexpr gpuThresholdId gpuTLimitHardshutDownThresholdId{4};
+
+void readThermalParameters(
+    uint8_t eid, const std::vector<gpuThresholdId>& ids,
+    mctp::MctpRequester& mctpRequester,
+    const std::function<void(uint8_t, std::vector<int32_t>)>& callback);
diff --git a/src/nvidia-gpu/meson.build b/src/nvidia-gpu/meson.build
index 57bab75..4044eb3 100644
--- a/src/nvidia-gpu/meson.build
+++ b/src/nvidia-gpu/meson.build
@@ -5,6 +5,7 @@
     'NvidiaGpuMctpVdm.cpp',
     'NvidiaGpuSensor.cpp',
     'NvidiaGpuSensorMain.cpp',
+    'NvidiaGpuThresholds.cpp',
     'OcpMctpVdm.cpp',
 )
 
diff --git a/src/nvidia-gpu/tests/NvidiaGpuSensorTest.cpp b/src/nvidia-gpu/tests/NvidiaGpuSensorTest.cpp
index c630ffa..9455326 100644
--- a/src/nvidia-gpu/tests/NvidiaGpuSensorTest.cpp
+++ b/src/nvidia-gpu/tests/NvidiaGpuSensorTest.cpp
@@ -491,6 +491,156 @@
     EXPECT_EQ(result, EINVAL); // Should indicate error for invalid data size
 }
 
+// Tests for GpuMctpVdm::encodeReadThermalParametersRequest function
+TEST_F(GpuMctpVdmTests, EncodeReadThermalParametersRequestSuccess)
+{
+    const uint8_t instanceId = 5;
+    const uint8_t sensorId = 1;
+    std::array<uint8_t, sizeof(gpu::ReadThermalParametersRequest)> buf{};
+
+    int result =
+        gpu::encodeReadThermalParametersRequest(instanceId, sensorId, buf);
+
+    EXPECT_EQ(result, 0);
+
+    gpu::ReadThermalParametersRequest request{};
+    std::memcpy(&request, buf.data(), sizeof(request));
+
+    EXPECT_EQ(request.hdr.msgHdr.hdr.pci_vendor_id,
+              htobe16(gpu::nvidiaPciVendorId));
+    EXPECT_EQ(request.hdr.msgHdr.hdr.instance_id &
+                  ocp::accelerator_management::instanceIdBitMask,
+              instanceId & ocp::accelerator_management::instanceIdBitMask);
+    EXPECT_NE(request.hdr.msgHdr.hdr.instance_id &
+                  ocp::accelerator_management::requestBitMask,
+              0);
+    EXPECT_EQ(request.hdr.msgHdr.hdr.ocp_accelerator_management_msg_type,
+              static_cast<uint8_t>(gpu::MessageType::PLATFORM_ENVIRONMENTAL));
+
+    // Verify request data
+    EXPECT_EQ(request.hdr.command,
+              static_cast<uint8_t>(
+                  gpu::PlatformEnvironmentalCommands::READ_THERMAL_PARAMETERS));
+    EXPECT_EQ(request.hdr.data_size, sizeof(sensorId));
+    EXPECT_EQ(request.sensor_id, sensorId);
+}
+
+// Tests for GpuMctpVdm::decodeReadThermalParametersResponse function
+TEST_F(GpuMctpVdmTests, DecodeReadThermalParametersResponseSuccess)
+{
+    // Create a mock successful response
+    std::array<uint8_t, sizeof(gpu::ReadThermalParametersResponse)> buf{};
+
+    gpu::ReadThermalParametersResponse response{};
+    ocp::accelerator_management::BindingPciVidInfo headerInfo{};
+    headerInfo.ocp_accelerator_management_msg_type = static_cast<uint8_t>(
+        ocp::accelerator_management::MessageType::RESPONSE);
+    headerInfo.instance_id = 5;
+    headerInfo.msg_type =
+        static_cast<uint8_t>(gpu::MessageType::PLATFORM_ENVIRONMENTAL);
+
+    gpu::packHeader(headerInfo, response.hdr.msgHdr.hdr);
+
+    // Populate response data
+    response.hdr.command = static_cast<uint8_t>(
+        gpu::PlatformEnvironmentalCommands::READ_THERMAL_PARAMETERS);
+    response.hdr.completion_code = static_cast<uint8_t>(
+        ocp::accelerator_management::CompletionCode::SUCCESS);
+    response.hdr.reserved = 0;
+    response.hdr.data_size = htole16(sizeof(int32_t));
+
+    // Set a threshold value of 85°C (85 * 256 = 21760)
+    response.threshold = htole32(21760);
+
+    std::memcpy(buf.data(), &response, sizeof(response));
+
+    // Test decoding
+    ocp::accelerator_management::CompletionCode cc{};
+    uint16_t reasonCode{};
+    int32_t threshold{};
+
+    int result = gpu::decodeReadThermalParametersResponse(
+        buf, cc, reasonCode, threshold);
+
+    EXPECT_EQ(result, 0);
+    EXPECT_EQ(cc, ocp::accelerator_management::CompletionCode::SUCCESS);
+    EXPECT_EQ(reasonCode, 0);
+    EXPECT_EQ(threshold, 21760);
+}
+
+TEST_F(GpuMctpVdmTests, DecodeReadThermalParametersResponseError)
+{
+    std::array<uint8_t,
+               sizeof(ocp::accelerator_management::CommonNonSuccessResponse)>
+        buf{};
+
+    // Populate error response data
+    ocp::accelerator_management::CommonNonSuccessResponse errorResponse{};
+    ocp::accelerator_management::BindingPciVidInfo headerInfo{};
+    headerInfo.ocp_accelerator_management_msg_type = static_cast<uint8_t>(
+        ocp::accelerator_management::MessageType::RESPONSE);
+    headerInfo.instance_id = 5;
+    headerInfo.msg_type =
+        static_cast<uint8_t>(gpu::MessageType::PLATFORM_ENVIRONMENTAL);
+
+    gpu::packHeader(headerInfo, errorResponse.msgHdr.hdr);
+
+    errorResponse.command = static_cast<uint8_t>(
+        gpu::PlatformEnvironmentalCommands::READ_THERMAL_PARAMETERS);
+    errorResponse.completion_code = static_cast<uint8_t>(
+        ocp::accelerator_management::CompletionCode::ERR_NOT_READY);
+    errorResponse.reason_code = htole16(0x5678);
+
+    std::memcpy(buf.data(), &errorResponse, sizeof(errorResponse));
+
+    // Test decoding
+    ocp::accelerator_management::CompletionCode cc{};
+    uint16_t reasonCode{};
+    int32_t threshold{};
+
+    int result = gpu::decodeReadThermalParametersResponse(
+        buf, cc, reasonCode, threshold);
+
+    EXPECT_EQ(result, 0);
+    EXPECT_EQ(cc, ocp::accelerator_management::CompletionCode::ERR_NOT_READY);
+    EXPECT_EQ(reasonCode, 0x5678);
+}
+
+TEST_F(GpuMctpVdmTests, DecodeReadThermalParametersResponseInvalidSize)
+{
+    // Create a mock response with invalid data_size
+    std::array<uint8_t, sizeof(gpu::ReadThermalParametersResponse)> buf{};
+
+    gpu::ReadThermalParametersResponse response{};
+    ocp::accelerator_management::BindingPciVidInfo headerInfo{};
+    headerInfo.ocp_accelerator_management_msg_type = static_cast<uint8_t>(
+        ocp::accelerator_management::MessageType::RESPONSE);
+    headerInfo.instance_id = 5;
+    headerInfo.msg_type =
+        static_cast<uint8_t>(gpu::MessageType::PLATFORM_ENVIRONMENTAL);
+
+    gpu::packHeader(headerInfo, response.hdr.msgHdr.hdr);
+
+    response.hdr.command = static_cast<uint8_t>(
+        gpu::PlatformEnvironmentalCommands::READ_THERMAL_PARAMETERS);
+    response.hdr.completion_code = static_cast<uint8_t>(
+        ocp::accelerator_management::CompletionCode::SUCCESS);
+    response.hdr.reserved = 0;
+    response.hdr.data_size = htole16(2); // Invalid - should be sizeof(int32_t)
+    response.threshold = htole32(21760);
+
+    std::memcpy(buf.data(), &response, sizeof(response));
+
+    // Test decoding
+    ocp::accelerator_management::CompletionCode cc{};
+    uint16_t reasonCode{};
+    int32_t threshold{};
+
+    int result = gpu::decodeReadThermalParametersResponse(
+        buf, cc, reasonCode, threshold);
+
+    EXPECT_EQ(result, EINVAL); // Should indicate error for invalid data size
+}
 } // namespace gpu_mctp_tests
 
 int main(int argc, char** argv)