From Disaggregated RAN to Unified Automation: Ericsson’s EIAP Expansion Signals the Next Phase of AI-Native Networks

From Disaggregated RAN to Unified Automation: Ericsson’s EIAP Expansion Signals the Next Phase of AI-Native Networks

The Long Road to Intelligent Networks

The telecom industry’s journey toward AI-native networks has been anything but linear. What began as a push for vendor diversity and cost reduction through Open RAN disaggregation in the late 2010s evolved into a sophisticated ecosystem of intelligent controllers, machine learning models, and now, unified automation platforms spanning the entire network. On July 6, 2026, Ericsson announced the expansion of its Ericsson Intelligent Automation Platform (EIAP) to encompass core network automation, extending its proven open, multi-vendor capabilities from the Radio Access Network (RAN) domain.

This development marks a pivotal moment. For years, operators have grappled with siloed operations—RAN intelligence via the RAN Intelligent Controller (RIC) on one side, core network management on the other. Ericsson’s move unifies these domains under a single, AI-driven framework, promising streamlined operations, optimized costs, and resilient performance.

Roots in Open RAN Disaggregation

To appreciate today’s news, we must trace back to the origins of Open RAN. The O-RAN Alliance, formed around 2018, championed standardized interfaces that broke the proprietary lock-in of traditional RAN vendors. This disaggregation allowed operators to mix and match components: radios from one vendor, distributed units (DUs) from another, and centralized units (CUs) running on commercial off-the-shelf hardware.

Early Open RAN deployments focused on interoperability and cost savings. By 2023-2024, attention shifted to intelligence. The introduction of the Near-Real-Time RIC (Near-RT RIC) and Non-Real-Time RIC (Non-RT RIC) enabled xApps and rApps—small applications that optimize RAN behavior using AI/ML. Operators like Rakuten and Dish began trialing these in live networks, demonstrating gains in energy efficiency and traffic steering.

Ericsson, initially cautious about full Open RAN, embraced its principles through its Cloud RAN portfolio and contributions to O-RAN standards. The company’s early AI work centered on the RAN, with features like AI-powered beamforming and predictive maintenance embedded in its radios and baseband.

The Rise of AI in the RAN

Parallel to Open RAN, the AI-RAN concept gained traction. NVIDIA’s Aerial platform and partnerships (including with Nokia, as previously covered on this blog) highlighted GPU-accelerated baseband processing that doubles as an AI inference engine. The AI-RAN Alliance, now with over 130 members, has driven blueprints for energy-efficient, AI-native 5G-Advanced and 6G networks.

Ericsson contributed significantly here, launching AI-ready radios with neural network accelerators and participating in the AI-RAN Alliance. Its Intelligent Automation Platform initially targeted RAN domains, leveraging open interfaces for multi-vendor orchestration. rApps could dynamically adjust parameters across heterogeneous networks, reducing OPEX while improving QoS.

Live trials, such as those with KDDI in Japan using Samsung equipment, showed AI-driven RAN optimizations delivering up to 52% speed gains. These successes validated the RIC model but exposed a gap: core networks remained largely manual or vendor-specific.

Bridging RAN and Core: The EIAP Expansion

Ericsson’s July 6, 2026 announcement addresses this gap directly. The EIAP now extends its open, multi-vendor automation to the core, creating a unified platform for Communication Service Providers (CSPs). Key capabilities include:

  • End-to-End Orchestration: AI models that correlate RAN events with core signaling for proactive optimization.
  • Autonomous Networks Vision: Alignment with TM Forum and 3GPP standards for zero-touch operations.
  • Multi-Vendor Support: Leveraging O-RAN interfaces while extending to core elements like AMF, SMF, and UPF.

This evolution empowers operators to treat the entire network as a single intelligent system. For instance, congestion detected in the RAN can trigger core-side policy adjustments automatically, all orchestrated through AI.

Historical Parallels and Industry Momentum

This mirrors earlier industry shifts. Just as SDN and NFV disaggregated the core in the 2010s, Open RAN did the same for the edge. Intelligence followed disaggregation: first in the RAN via RIC, now across domains.

Dell’Oro Group’s recent forecasts underscore the stakes, projecting cumulative AI RAN revenue reaching $35 billion by 2030 amid broader Open RAN integration. Ericsson’s platform positions it to capture value beyond RAN hardware, moving into software and services for autonomous networks.

Competitors are not idle. Nokia’s deep ties with NVIDIA for GPU-based AI-RAN, Samsung’s path to AI optimizers, and startups like DeepSig advancing AI-native air interfaces all point to an industry converging on software-defined, AI-augmented infrastructure.

Implications for Operators and the Road to 6G

For Open RAN operators, this expansion lowers barriers to full-network AI adoption. Existing RIC investments can extend upward, avoiding new siloes. Energy efficiency, a perennial concern, improves as AI balances load across RAN and core.

Looking ahead to 6G, where AI-native design is foundational, unified platforms like EIAP provide the blueprint. Field trials expected in 2026-2027 will validate gains in resilience and new service enablement, from edge AI to network slicing for humanoid robotics or industrial automation.

Challenges remain: ensuring data privacy in cross-domain AI, standardizing interfaces beyond O-RAN, and upskilling workforces. Yet the trajectory is clear—from fragmented hardware to intelligent, open software ecosystems.

Ericsson’s move is not revolutionary in isolation but evolutionary in context: the logical next step in a decade-long march toward networks that think, adapt, and optimize themselves.

Sources

Related Posts

South Korea's Hyper AI Network Initiative: From Open RAN Disaggregation to Physical AI in Industry

South Korea's new $11.6M government project tests AI-RAN and 5G for industrial robotics, building on years of Open RAN and AI evolution.

Nokia Launches Industry-First Commercial AI-RAN Platform: 5 Things It Means for Open RAN Operators

Nokia's July 15, 2026 AI-RAN platform announcement delivers GPU-powered intelligence, Open RAN compliance, and massive spectral gains—here's what operators need to know.