DeepSig Demonstrates 5G AI-Native Massive MIMO Open RAN Performance Improvement Milestone with Intel’s FlexRAN Reference Architecture

Demonstration Showed Faster and More Accurate Massive MIMO Performance for 5G Open RAN

ARLINGTON, Va.--()--DeepSig, experts in artificial intelligence (AI) and machine learning (ML) innovation in wireless communications, today announced that it hit a joint milestone in Open radio access networks (RAN) 5G Massive MIMO (mMIMO) performance leveraging Intel’s FlexRAN reference architecture.

Working with Intel, DeepSig has demonstrated artificial intelligence dramatically improves the performance and reduces power consumption of 5G mMIMO Open RAN. DeepSig’s AI-Native OmniPHY® 5G software integrates into Intel FlexRAN Layer 1 reference software and operates fully within the 5G Open Distributed Unit (O-DU).

Massive MIMO demonstration with AI / ML algorithms, when compared to classic RAN receiver algorithms, showed up to 5X channel estimation speed improvements and accuracy was improved by up to 4X. OmniPHY 5G provides Open RAN operators an alternative mMIMO AI solution in their existing DU hardware with lower TCO, higher data rates and greater spectral efficiency, especially under high mobility conditions where speed and latency are critical. OmniPHY 5G was demonstrated on 3rd Gen Intel Xeon Scalable processors and 4th Gen Intel Xeon Scalable processors, leveraging Intel Deep Learning Boost (Intel DL Boost) acceleration.

DeepSig, in January 2023 delivered its first operational release of OmniPHY 5G software compiled into Intel FlexRAN L1 reference software to select Open RAN ecosystem partners. Open RAN Network Equipment Providers and Technology Providers can now test and integrate DeepSig’s AI neural receiver into their 5G Open vRAN solutions to benefit mobile operators and communications service providers.

DeepSig is set to release its mMIMO upgrade for OmniPHY 5G in the coming months to make Open RAN performance more competitive in dense multi-user macro-cell environments, where both spectral efficiency and energy efficiency are critical to operators.

“We are on track for productizing high-value, AI-native solutions for the Open RAN ecosystem,” said Jim Shea, DeepSig CEO. “Our experienced software team delivered the wireless industry’s first 5G AI-native over-the-air call in November 2021, achieved substantial throughput gains and power savings in 5G MIMO in 2022, and delivered the first 5G neural-receiver operational release to partners in January 2023. We’ve had very productive, close collaboration with Intel’s FlexRAN software team for over 3 years. We are excited for partners and operators to deploy DeepSig’s AI-native software with FlexRAN in their 5G vRAN products this year.”

“Intel’s ecosystem collaborations continue to demonstrate AI innovation will deliver considerable value to wireless operators as they deploy virtualized RAN architectures,” said Cristina Rodriguez, Vice President and General Manager, Wireless Access Networking Division at Intel. “DeepSig has been an early AI partner to leverage the Intel FlexRAN reference architecture as the foundation for-based implementation for Layer 1 and above. FlexRAN reference architecture enables many use cases in 5G commercial networks, private networks and the AI-native path beyond 5G.”

Attendees to Mobile World Congress 2023 in Barcelona can visit DeepSig in Hall 2, L14MR to learn more about their leading AI-native products for wireless systems.

About DeepSig, Inc.

DeepSig, Inc. is a venture-backed and product-centric technology company developing revolutionary wireless processing software solutions using cutting edge machine learning techniques to transform baseband processing, wireless sensing, and other key wireless applications. Known as “deep learning,” a proven technology in vision and speech processing now accelerates 5G network performance, capacity, operational efficiency, and the customer experience. For more information, visit


Summer Wilcox, DeepSig Inc., +1 757.870.9037 and



Summer Wilcox, DeepSig Inc., +1 757.870.9037 and