Ericsson releases AI-ready wireless, antenna equipment, and AI RAN software to empower future networks

[Communications Industry Network News] As AI-powered devices accelerate their global adoption, the importance of mobile networks seamlessly meeting new performance demands becomes increasingly evident. To address these evolving needs driven by multimodal AI and augmented reality (AR) use cases, mobile networks require stronger uplink performance. Before the 2026 Mobile World Congress (MWC), Ericsson announced a new suite of wireless, antenna, and AI Radio Access Network (AI RAN) software solutions.

These new products will help operators monetize AI devices through differentiated connectivity and introduce new services. The series includes AI-ready wireless devices equipped with neural network accelerators and Ericsson Silicon, which significantly enhances AI inference capabilities for Massive MIMO wireless equipment. Neural network accelerators are the core of programmable matrices, integrated into Ericsson’s multi-core architecture, optimized for AI and machine learning.

The new product portfolio is built around three pillars:

1. Ten AI-ready wireless products

Designed to improve downlink efficiency and achieve uplink performance breakthroughs, these next-generation Massive MIMO and remote radio units include:

• Performance breakthroughs such as high-power FDD Massive MIMO AIR3286, integrated TDD and FDD Massive MIMO AIR3211, and support for 8 receive channels Radio4891 and scalable Radio4458, all enhancing uplink performance for AI and AR applications.

• High-power tri-band remote radio units: Radio4488 and Radio4464, designed for network integration and RAN sharing.

• TDD Massive MIMO enhancements: AIR3267 achieves 600MHz instantaneous bandwidth at just 13kg; AIR6492 offers 480W output power and 256 antenna elements, delivering a performance leap.

2. RAN software

A series of new AI-powered features aimed at improving efficiency and reliability while meeting the latency requirements of AI applications.

• New AI-powered features include AI-managed beamforming, AI-enabled outdoor positioning, and advanced AI models for real-time coverage prediction, complementing Ericsson’s market-leading AI-native link adaptation software.

• To support CSPs in meeting the low latency and stable uplink rates required by AI and AR applications, new tools such as Latency Prioritized Scheduler and Low Latency Mobility have been added, boosting response speeds by up to seven times and providing better, more reliable connections.

In summary, these AI software enhancements increase efficiency and intelligence, enabling new services previously unattainable.

3. Five high-performance antenna products

Interleaved and passive antennas improve spectrum utilization and simplify site design. The new high-efficiency passive antennas, based on the trio net design, maximize uplink performance, carrier aggregation, and spectrum efficiency. The expanded interleaved AIR product lineup adds three new configurations to the existing two, fully supporting flexible TDD and FDD Massive MIMO deployments.

Ericsson’s Head of Network Strategy and Product Management, Morten Lerner, said: “As AI changes traffic patterns and raises consumer expectations, networks must deliver precise performance where and when it’s needed. At Ericsson, we are committed to providing this performance. We are embedding AI solutions across our entire product portfolio and launching AI RAN software that brings revolutionary improvements in spectrum efficiency. Now, with the introduction of neural network accelerators into our leading Massive MIMO portfolio, we are taking the final step toward fully AI-enabled products.”

The latest additions to Ericsson’s product portfolio deepen AI integration into the RAN, enabling real-time optimization of performance and energy use. With these AI-driven features and robust hardware lineup, CSPs will be able to deliver stable, differentiated user experiences. The portfolio aims to leverage the growing demand for AI-enabled devices by providing exceptional uplink performance and service differentiation. Additionally, AI-driven hardware and software efficiencies help reduce total cost of ownership (TCO) and accelerate value realization.

Iain Milligan, Network Development and Infrastructure Director at Vodafone Three UK, said: “By integrating Vodafone and Three’s networks into a high-performance 5G infrastructure supported by Ericsson RAN, Vodafone Three is leading a new era of connectivity—pioneering next-generation technologies and paving the way for AI-driven innovation. We are embedding AI throughout the network to reduce operational complexity, improve energy efficiency, and deliver faster, more reliable connections to customers across the country.”