Verizon’s Nationwide EIC Rollout: The SMO/RIC Layer Becomes the Real AI-RAN Battleground

On November 3, 2026, Verizon and Ericsson confirmed what many in the Open RAN community had been anticipating: a nationwide deployment of the Ericsson Intelligent Controller (EIC) across Verizon’s US network, with initial deployments beginning before the end of 2026. The announcement, carried by Telecoms.com, MarketScreener, Telecompaper, and TelecomTV, is more than another AI-RAN press release. It is a concrete signal that the service management and orchestration (SMO) and RAN Intelligent Controller (RIC) layer—the control plane for automation—is where the AI-RAN battle is now being fought. For engineers building and operating Open RAN systems, the Verizon-Ericsson deal offers a rare look at how a tier-one operator intends to turn rApps from lab curiosities into production-grade network functions.

What Verizon and Ericsson Are Actually Deploying

The core of the announcement is the EIC, a key component of Ericsson’s Intelligent Automation Platform (EIAP). EIAP is Ericsson’s open network management and automation platform, designed for multivendor and multitechnology environments. It supports Ericsson’s 4G and 5G RAN portfolio as well as core networks. The EIC handles the execution of AI-native RAN automation applications—rApps—and, critically, supports open standards that allow Verizon to develop and run its own rApps.

Verizon’s stated plan is to deploy EIC nationwide beginning this year, with initial deployments set to begin before the end of 2026. The operator intends to test and deploy new rApps developed either internally or by Ericsson, including the Cell Anomaly Detector, Root Cause Explainer, and Anomaly General Optimizer. These are not trivial applications. Cell Anomaly Detector aims to identify anomalous cell behavior automatically; Root Cause Explainer seeks to diagnose the underlying cause of performance degradations; and Anomaly General Optimizer is intended to optimize network performance in response to detected anomalies. Together, they represent a shift from static, rule-based SON functions toward AI-driven, closed-loop automation at the non-real-time RIC level.

Verizon has already demonstrated tangible results. According to the company, it has deployed fully autonomous optimization rApps that boosted downlink and uplink throughput by 8.9% and 8%, respectively. It has also implemented an rApp that eliminates the need for manual intervention in the event of a handover failure from LTE to 5G standalone (SA) for fixed-wireless access (FWA) devices. Matt Butz, VP of network planning at Verizon, said, “Using the Ericsson Intelligent Controller, and Ericsson rApps, in a live environment, we demonstrated AI cell shaping of Massive MIMO and traffic steering with significant gains in throughput and efficiency. Verizon’s ability to self-develop rApps on this open platform will help usher in a new level of automation as we prepare for 6G.”

These figures—8.9% downlink and 8% uplink—are notable because they come from a live environment, not a lab. They suggest that AI-driven non-real-time control can deliver meaningful spectral efficiency gains even before the industry moves to more exotic AI-RAN architectures like AI-and-RAN (co-hosting RAN and external AI workloads on shared GPU infrastructure) or AI-on-RAN (edge AI services for subscribers). The Verizon-Ericsson rollout is squarely in the AI-for-RAN category: the deeper integration of AI and ML for RAN automation and optimization.

The SMO/RIC Layer: Where the Intelligence Actually Lives

The Verizon announcement is best understood in the context of a broader market shift. According to SNS Telecom & IT, annual spending on AI-RAN automation and optimization software and services is expected to grow at a CAGR of 9% over the next four years, reaching US$1.1 billion by the end of 2030. The wider RAN automation market—including traditional SON/proprietary network-level automation, AI-for-RAN upgrades, Open RAN SMO/RIC and d/x/rApps, RF planning tools, and automated test/measurement solutions—is expected to grow at a CAGR of approximately 4% during the same period.

But the aggregate numbers obscure a more important trend: the SMO/RIC segment is growing far faster than the rest. SNS Telecom & IT projects that global spending on SMO, RIC, and d/x/rApps will grow at a CAGR of approximately 40% between 2026 and 2030, driven by RAN modernization and hybrid/Open RAN infrastructure rollouts by brownfield operators. The Open RAN automation market will represent more than $660 million in annual investments by the end of 2030 as SMO/Non-RT RIC platforms and their associated application ecosystems gain traction in brownfield networks.

This is the key architectural insight. The AI-RAN conversation often focuses on the radio unit and the distributed unit—GPUs in the cell tower, AI-native baseband, and the like. But the Verizon-Ericsson deal shows that the most commercially mature and immediately deployable AI-RAN capability is at the non-real-time RIC and SMO layer. The EIC is essentially a non-real-time RIC that hosts rApps, which operate on timescales of seconds to minutes, not milliseconds. These rApps can perform cell shaping, traffic steering, anomaly detection, and root-cause analysis without requiring changes to the baseband processing or the radio hardware.

This matters because it sidesteps the biggest obstacle to AI-RAN adoption: the reluctance of DU/CU vendors to open their baseband platforms to third-party xApps. As SNS Telecom & IT notes, reluctance among DU/CU vendors has held back adoption of xApps and the Near-RT RIC. The EIC, by contrast, is an open platform that supports rApps from multiple sources, including the operator itself. Verizon’s ability to self-develop rApps is a direct benefit of this openness.

The E3 Interface and the Latency Wall

Of course, non-real-time rApps have their limits. They operate above the 10-millisecond latency boundary that defines the Near-RT RIC. For lower-layer performance optimization, geolocation, integrated sensing, spectrum sharing, and interference cancellation in 5G and future 6G deployments, operators need closed-loop control at much faster timescales.

This is where the recently proposed E3 interface comes in. E3 provides a standardized mechanism for dApps to directly interface with Open RAN baseband nodes, enabling closed-loop control beyond the 10-millisecond latency limits of xApps. The E3 interface is gaining growing interest as a way to unlock lower-layer optimization without requiring full Near-RT RIC adoption. Some vendors are pursuing an alternative: offering proprietary AI-enabled, real-time optimization capabilities as software upgrades to existing RAN platforms through subscription-based licensing models.

The Verizon-Ericsson deal does not mention E3 or dApps, but it is part of the same trajectory. Verizon’s stated goal is to reach Level 4 cognitive automation on its network. According to TelecomTV, many operators aim to achieve L4 autonomy for high-impact scenarios such as energy savings, traffic optimization, fault management, service assurance, and special event management by 2030. The pursuit of highly autonomous network operations, coupled with the multi-vendor interoperability, scalability, and latency-related limitations of legacy SON solutions, is prompting operators to embrace Open RAN automation with standards-based components that enable greater RAN programmability and AI/ML-driven automation.

Verizon’s EIC deployment is a step toward that L4 goal. The rApps it is deploying—Cell Anomaly Detector, Root Cause Explainer, Anomaly General Optimizer—are exactly the kind of high-impact, non-real-time functions that can be automated at the SMO/RIC layer. They do not require sub-10-millisecond control loops. They require data, analytics, and policy-based decision-making, which is what the non-real-time RIC is designed for.

The Competitive Landscape: Why This Matters for Open RAN

The Verizon-Ericsson announcement is significant not just for what it says about Verizon, but for what it says about the competitive dynamics of the AI-RAN market. Ericsson has secured a nationwide deployment with a major US operator for its EIC platform. This is a vote of confidence in Ericsson’s approach to open, standards-based automation. But it also raises questions about the balance of power in the Open RAN ecosystem.

On one hand, the EIC’s support for open standards and Verizon’s ability to develop its own rApps demonstrate that the platform is genuinely open. Verizon is not locked into Ericsson-only applications. It can build its own rApps and run them on the EIC. This is a meaningful step toward the vision of an open, multi-vendor RAN automation ecosystem.

On the other hand, the fact that a tier-one operator is deploying a single vendor’s intelligent controller nationwide concentrates a significant amount of control in one vendor’s platform. The EIC becomes the point through which all rApps—whether from Ericsson, Verizon, or third parties—must flow. This is not necessarily a problem, but it is a dynamic that Open RAN advocates should watch closely. The SMO/RIC layer is becoming the new control point for the RAN, and whoever controls that layer has significant influence over the pace and direction of innovation.

This dynamic is reflected in the broader market. SNS Telecom & IT notes that while reluctance among DU/CU vendors has held back adoption of xApps and the Near-RT RIC, there is growing interest in the E3 interface as a way to bypass the Near-RT RIC and enable direct control of baseband nodes. If E3 gains traction, it could shift the balance of power again, giving dApp developers direct access to the baseband and reducing the importance of the Near-RT RIC. But for now, the non-real-time RIC—and the SMO layer more broadly—is where the action is.

The Road to 6G and the Role of Digital Twins

The Verizon-Ericsson deal is also a bet on the future. The launch of the first 6G networks—which are being standardized as AI-native systems from the outset—around 2030 is expected to significantly increase the scope for hardware-agnostic, software-led innovation and take the industry closer to fully autonomous network operations. RAN digital twins are expected to play an increasingly important role, allowing operators to validate AI-enabled automation before deploying it in live networks.

Verizon’s deployment of EIC and rApps is a way to build the operational experience and data foundation needed for that future. By deploying AI-native automation now, Verizon is learning how to manage AI-driven networks, how to validate rApps, and how to integrate them into its operational processes. This experience will be invaluable as the industry moves toward 6G.

It is also worth noting that the EIC supports Ericsson’s entire portfolio of 4G and 5G RAN, as well as core networks. This means Verizon can use the same platform to automate across multiple generations of technology, simplifying operations and reducing the need for separate automation silos. This is a practical benefit that should not be overlooked.

What Engineers Should Take Away

For engineers working on Open RAN and AI-RAN, the Verizon-Ericsson announcement offers several practical lessons.

First, the SMO/RIC layer is where AI-RAN is being deployed today. The non-real-time RIC, with its rApps, is the most commercially mature facet of the AI-RAN movement. It does not require new hardware, new spectrum, or new baseband processing. It leverages existing data and existing interfaces to deliver measurable gains. The 8.9% downlink and 8% uplink throughput improvements Verizon reported are real and achievable with current technology.

Second, openness matters. Verizon’s ability to self-develop rApps on the EIC is a direct result of the platform’s support for open standards. Operators should demand this kind of openness from their vendors. The ability to build and run their own applications is a key lever for differentiation and innovation.

Third, the latency wall is real. Non-real-time rApps are powerful, but they cannot address all use cases. For lower-layer optimization, integrated sensing, and interference cancellation, operators will need faster control loops. The E3 interface and dApps are promising developments, but they are not yet widely deployed. Engineers should watch this space closely.

Fourth, the path to L4 autonomy is incremental. Verizon’s deployment of specific rApps for specific use cases—handover failure, cell anomaly detection, root cause analysis—is a step-by-step approach. It is not a big-bang transformation. This is the right way to build confidence and demonstrate value.

Conclusion

The Verizon-Ericsson nationwide EIC deployment is a milestone in the evolution of AI-RAN. It shows that the SMO/RIC layer is not just a theoretical concept but a practical, deployable platform for AI-driven automation. It demonstrates that operators can achieve meaningful gains—8.9% downlink, 8% uplink—with current technology. And it highlights the importance of openness in the automation platform.

As the industry moves toward 6G and fully autonomous networks, the SMO/RIC layer will only become more important. The battle over where intelligence lives—in the radio, in the baseband, in the RIC, or in the cloud—is far from over. But the Verizon-Ericsson deal suggests that for now, the non-real-time RIC is the place to be. Engineers should take note.

Sources

• Verizon ramps up AI-RAN activity with Ericsson rApp rollout
• Verizon and Ericsson team up to integrate AI into RAN automation
• Verizon commences roll-out of AI RAN optimisations through Ericsson Intelligent Controller
• What’s up with… Verizon, Deutsche Telekom, ETSI, The Future of RAN
• AI-enabled RAN automation and optimisation a US$1.1 billion market, says SNS Telecom & IT
• SMO/RIC-based RAN automation gains momentum (Reader Forum)