It’s Time for Data Management Companies to Take a More Proactive Role in Helping Telcos with RAN Automation

Throughout the duration of MWC Barcelona 2025, discussions around advancements in Artificial Intelligence (AI) continued to dominate the halls, even more than in 2024 as the industry now looks more at the implementation of the technology, rather than its availability. Within the Radio Access Network (RAN) domain, many of the announcements were related to natively integrating AI into the automation platforms, including the development of AI agents and vertical Large Language Models (LLMs) for telco operations.

Currently, telcos are notorious for having vast amounts of data that they are unable to utilize due to their unstructured or semi-structured nature, with most of the structure data found within the Business Support System (BSS) domain and varied levels of structure for the Operational Support System (OSS) depending on the telco. With RAN automation rapidly growing in importance as telcos aim to reach Level 4 automation as defined by TM Forum, telcos must strive to greatly improve the management and analytics of RAN data to extract the maximum benefits of the various AI technologies they integrate.

In March 2025, the AI-RAN Alliance launched its Data-for-AI initiative, which is focused on defining a structured approach to data management and data collection pipelines in order to ensure data availability for AI performance benchmarking and further research. This initiative is designed for Research & Development (R&D) purposes, rather than for commercial applications. Furthermore, Pure Storage also announced that it had joined the AI-RAN Alliance, aiming to help the industry with better managing the increased data traffic, diverse service requirements, and rapid scalability demands necessary when integrating AI more natively into RAN automation platforms, such as Radio Intelligent Controller (RIC) and Service Management & Orchestration (SMO).

Data management is a core aspect of RAN automation technologies, as operators use them to gain better insights into network performance, where new RAN deployments must be made, or where errors occur in their network to resolve issues rapidly, for example (for more details on the use cases for RAN automation platforms, please see ABI Research’s 5G RAN Automation Platforms: Use Cases for Network Operations report (AN-5972). Currently, deployed RAN automation platforms such as legacy Self-Organizing Networks (SON), SMO, and RIC are not integrated with the automation technologies within the core network, and the data management technologies remain distinct from a vendor’s respective OSS/BSS data management solutions, and require some level of configuration to individual telco-specific data structures. Furthermore, new offerings to customers such as network slices exacerbate this complexity as the telco must constantly be aware of where the user is located, which cell site is providing their connection, and ensure a seamless cell handover to maintain the customers experience, which involves keeping a close eye both on the subscriber data and RAN network data simultaneously for many different users simultaneously. This results in achieving End-to-End (E2E) network automation, which is much more complex while these systems remain siloed.

This does not mean that the industry is ignorant about this reality. There are efforts to try to bridge this gap, with companies such as Huawei and ZTE already claiming that their customers have achieved Level 4 automation, and SoftBank’s recently announced Large Telecom Model (LTM) set to be integrated with its AITRAS solution for its upcoming deployment of an AI RAN architecture in Japan. However, there are many challenges due to the fragmented and legacy nature of telco data. Due to the telco industry’s struggles thus far, data management companies have a great opportunity to enter the market and act as abstractors or intermediaries between telcos and RAN automation vendors. Data management companies are well positioned to help telcos, and the entire industry, accelerate their data transformation as they already solve similar problems.

Companies such as Databricks, Snowflake, and Ocient already play a role in the OSS/BSS markets for data management as enhancers, rather than direct replacements. Expanding their portfolios to offer a truly E2E data solution for telcos, which includes RAN analytics and data, is the natural next step for these players.

The above is much simpler in theoretical terms than it would be in a commercial setting. Incumbents such as Samsung, Nokia, and Ericsson already have some form of data management and analytics portfolios for both OSS/BSS and for their own RAN automation solutions (whether proprietary or RIC and SMO) with specific data formats and models. Data platforms have the capability to map telco-specific data into the expected structure to make onboarding faster, cheaper, and more standardized. The data platforms, in layman’s terms, becomes the data fabric of network automation.

In order for this to become a commercial reality, ABI Research recommends that data management companies adopt the following strategies:

• Join and Participate in the AI-RAN Alliance: With this alliance having publicly declared a focus on data management, joining the alliance is a natural entry point for data management companies. Even though the initiative is tailored toward R&D activities, showcasing their expertise in data management in this area may start conversations around the benefits of working closer with data management companies in the RAN domain. Joining this alliance, as well as others such as TM Forum, allows for more visibility into how the industry is progressing, and offers the opportunity for more influence over the direction of the industry in their data strategies for automation.
• Conduct Trials for RAN Data Management with Existing Telco Customers: Given that existing telco customers already trust data management companies for OSS/BSS, running RAN-centric Proofs of Concept (PoCs) to extract value for newer technologies such as digital twins and predicting network performance allows data management companies to show their RAN domain expertise. This can be for a variety of use cases such as RAN energy savings in a multi-vendor environment, for example.
• Establish Partnerships with RAN Automation Vendors: Partnering with leading RAN automation vendors such as Nokia, Ericsson, and Samsung will be key in order to establish a successful go-to-market strategy. Acting as neutral data intermediaries will require working closely with these vendors to ensure that the data on the platform fits their necessary schemas for their automation solutions, and in the long term, can reduce the time necessary for integration when securing new deals with telco customers for the data management companies. While telco data remain fragmented currently, data management companies can, in partnership with RAN automation vendors, provide best practice case studies for how to become more data driven and better utilize their data while remaining as secure as possible. This is essential as achieving E2E network automation is not just a technology issue, but requires upskilling the work.

That said, there are also some challenges to address. First, incumbent vendors such as Ericsson and Nokia already have existing partnerships with hyperscalers such as Amazon Web Services (AWS) and Google Cloud, both of which are perfectly capable of providing data platforms themselves, as well as hosting and managing RAN functions. Google Cloud announced its RAN Guardian Agent solution during MWC 2025, which shows that hyperscalers are also making a move into the RAN market. However, data management companies are able to act as more neutral intermediaries without their own RAN or automation products, which must be highlighted when attempting to go to market.