diff --git a/inference-engine/src/gna_plugin/gna_graph_compiler.cpp b/inference-engine/src/gna_plugin/gna_graph_compiler.cpp
index d3b522d89e4..d5ececd00b6 100644
--- a/inference-engine/src/gna_plugin/gna_graph_compiler.cpp
+++ b/inference-engine/src/gna_plugin/gna_graph_compiler.cpp
@@ -501,12 +501,23 @@ void GNAGraphCompiler::PowerPrimitive(InferenceEngine::CNNLayerPtr layer) {
     auto input = layer->insData[0].lock();
 
     auto outputs = *layer->outData.begin();
-
-    uint32_t num_rows_in = FROM_IR_DIM(input, 1);
-    uint32_t num_columns_in = FROM_IR_DIM(input, 2);
+    uint32_t num_rows_in = InferenceEngine::details::product(begin(input->getDims()), end(input->getDims()));
+    uint32_t num_columns_in = 1;
     uint32_t num_rows_out = num_rows_in;
     uint32_t num_padding = ALIGN(num_rows_in, 8) - num_rows_in;
 
+    if (input->getDims().size() > 2 || input->getDims()[0] >= 8) {
+        for (size_t index_divide = 8; index_divide > 0; index_divide--) {
+            if (num_rows_in % index_divide == 0) {
+                num_rows_in /= index_divide;
+                num_columns_in = index_divide;
+                break;
+            }
+        }
+        num_rows_out = num_rows_in;
+        num_padding = ALIGN(num_rows_in, 8) - num_rows_in;
+    }
+
     size_t num_data_bytes_out = InferenceEngine::details::product(begin(outputs->getDims()), end(outputs->getDims()))
         * outputs->getPrecision().size();
 
diff --git a/inference-engine/tests/functional/plugin/gna/shared_tests_instances/subgraph_tests/broadcast_power.cpp b/inference-engine/tests/functional/plugin/gna/shared_tests_instances/subgraph_tests/broadcast_power.cpp
new file mode 100644
index 00000000000..ebd357ff5a8
--- /dev/null
+++ b/inference-engine/tests/functional/plugin/gna/shared_tests_instances/subgraph_tests/broadcast_power.cpp
@@ -0,0 +1,49 @@
+// Copyright (C) 2020 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+
+#include <vector>
+#include <string>
+#include <ie_precision.hpp>
+#include <subgraph_tests/broadcast_power.hpp>
+
+using namespace LayerTestsDefinitions;
+
+namespace {
+const std::vector<InferenceEngine::Precision> netPrecisions = {
+        InferenceEngine::Precision::FP32,
+        InferenceEngine::Precision::FP16,
+};
+
+const std::vector<std::map<std::string, std::string>> configs = {
+        {
+                {"GNA_DEVICE_MODE", "GNA_SW_EXACT"},
+                {"GNA_COMPACT_MODE", "NO"},
+                {"GNA_SCALE_FACTOR_0", "2048"},
+        }
+};
+
+const std::vector<std::vector<std::vector<size_t>>> input_shapes {
+        {{1, 8224}, {1, 257, 32}},
+        {{2, 8224}, {1, 257, 64}},
+        {{4, 8224}, {1, 257, 128}},
+        {{8, 128}, {8, 128}},
+        {{16, 128}, {16, 128}},
+        {{18, 128}, {18, 128}},
+        {{1, 16, 1, 128}, {1, 16, 1, 128}},
+        {{1, 8, 15, 128}, {1, 8, 15, 128}},
+        {{4, 4, 4, 4}, {4, 4, 4, 4}},
+        {{1, 4, 4, 128}, {1, 4, 4, 128}}
+        //TODO: needed add split over channels
+//        {{8, 8224}},
+};
+
+
+INSTANTIATE_TEST_CASE_P(PowerBroadcast, BroadcastPowerTest,
+        ::testing::Combine(
+        ::testing::ValuesIn(input_shapes),
+        ::testing::ValuesIn(netPrecisions),
+        ::testing::Values(CommonTestUtils::DEVICE_GNA),
+        ::testing::ValuesIn(configs)),
+        BroadcastPowerTest::getTestCaseName);
+
+} // namespace
diff --git a/inference-engine/tests/functional/plugin/shared/include/subgraph_tests/broadcast_power.hpp b/inference-engine/tests/functional/plugin/shared/include/subgraph_tests/broadcast_power.hpp
new file mode 100644
index 00000000000..d682d467da8
--- /dev/null
+++ b/inference-engine/tests/functional/plugin/shared/include/subgraph_tests/broadcast_power.hpp
@@ -0,0 +1,33 @@
+// Copyright (C) 2020 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#pragma once
+
+#include <tuple>
+#include <vector>
+#include <string>
+#include <memory>
+
+#include "functional_test_utils/layer_test_utils.hpp"
+#include "ngraph_functions/utils/ngraph_helpers.hpp"
+#include "ngraph_functions/builders.hpp"
+
+typedef std::tuple<
+        std::vector<std::vector<size_t>>,   // Input shapes
+        InferenceEngine::Precision,         // Network Precision
+        std::string,                        // Target Device
+        std::map<std::string, std::string>  //Configuration
+> BroadCastPowerTuple;
+
+namespace LayerTestsDefinitions {
+
+class BroadcastPowerTest : public testing::WithParamInterface<BroadCastPowerTuple>,
+                         virtual public LayerTestsUtils::LayerTestsCommon {
+public:
+    static std::string getTestCaseName(testing::TestParamInfo<BroadCastPowerTuple> obj);
+
+protected:
+    void SetUp() override;
+};
+}  // namespace LayerTestsDefinitions
diff --git a/inference-engine/tests/functional/plugin/shared/src/subgraph_tests/broadcast_power.cpp b/inference-engine/tests/functional/plugin/shared/src/subgraph_tests/broadcast_power.cpp
new file mode 100644
index 00000000000..3fd88279fce
--- /dev/null
+++ b/inference-engine/tests/functional/plugin/shared/src/subgraph_tests/broadcast_power.cpp
@@ -0,0 +1,52 @@
+// Copyright (C) 2020 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#include <debug.h>
+#include "common_test_utils/common_utils.hpp"
+#include "functional_test_utils/precision_utils.hpp"
+#include "functional_test_utils/skip_tests_config.hpp"
+#include "subgraph_tests/broadcast_power.hpp"
+
+namespace LayerTestsDefinitions {
+std::string BroadcastPowerTest::getTestCaseName(testing::TestParamInfo<BroadCastPowerTuple> obj) {
+    InferenceEngine::Precision netPrecision;
+    std::string targetDevice;
+    std::map<std::string, std::string> configuration;
+    std::vector<std::vector<size_t>> inputs_shapes;
+    std::tie(inputs_shapes, netPrecision, targetDevice, configuration) = obj.param;
+
+    std::ostringstream result;
+    result << "inputShape=" << CommonTestUtils::vec2str(inputs_shapes) << "_";
+    result << "netPRC=" << netPrecision.name() << "_";
+    result << "targetDevice=" << targetDevice << "_";
+    for (auto const& configItem : configuration) {
+        result << "_configItem=" << configItem.first << "_" << configItem.second;
+    }
+    return result.str();
+}
+
+void BroadcastPowerTest::SetUp() {
+    InferenceEngine::Precision netPrecision;
+    std::vector<std::vector<size_t>> inputs_shapes;
+    std::tie(inputs_shapes, netPrecision, targetDevice, configuration) = this->GetParam();
+    auto ngPrc = FuncTestUtils::PrecisionUtils::convertIE2nGraphPrc(netPrecision);
+
+    auto params = ngraph::builder::makeParams(ngPrc, {inputs_shapes[0]});
+    auto reshape_pattern = std::make_shared<ngraph::op::Constant>(ngraph::element::i64, ngraph::Shape{inputs_shapes[1].size()},
+            inputs_shapes[1]);
+    auto reshape = std::make_shared<ngraph::opset1::Reshape>(params[0], reshape_pattern, false);
+
+    auto const_mult2 = ngraph::builder::makeConstant<float>(ngPrc, {}, {-1.0f});
+    auto sum = ngraph::builder::makeEltwise(reshape, const_mult2, ngraph::helpers::EltwiseTypes::MULTIPLY);
+
+    auto reshape_pattern_2 = std::make_shared<ngraph::op::Constant>(ngraph::element::i64, ngraph::Shape{inputs_shapes[0].size()},
+            inputs_shapes[0]);
+    auto reshape_2 = std::make_shared<ngraph::opset1::Reshape>(sum, reshape_pattern_2, false);
+    function = std::make_shared<ngraph::Function>(reshape_2, params, "BroadcastPowerPass");
+}
+
+TEST_P(BroadcastPowerTest, CompareWithRefImpl) {
+    Run();
+};
+} // namespace LayerTestsDefinitions