Fujitsu demonstrated a bold, AI-driven approach to 5G at MWC Barcelona 2025, positioning its AI-RAN strategy as a core pillar for telcos seeking new ROI from next-generation networks. The company highlighted how artificial intelligence can optimize how GPU resources are allocated on the network, turning RAN functions into intelligent workloads, and how Open RAN and photonics underpin a scalable, open, high-capacity framework. In conversations at the event, Fujitsu’s Carlos Cordero, the Chief Technology Officer for Fujitsu Spain, described how the company bridges global research with customer needs to translate leading-edge technologies into practical telco benefits. This article expands on Fujitsu’s AI-led strategy, exploring the AI-RAN concept, its photonics and Open RAN foundations, the growing role of Private 5G, the transformative potential of 5G Standalone, and the broader market dynamics shaping telco investment, deployment, and long-term value realization.
AI-RAN and the Future of Resource Allocation
How AI-RAN Works
At the heart of Fujitsu’s strategy lies the AI-RAN concept, a forward-looking approach to distributing network hardware resources—specifically GPU servers—by applying artificial intelligence to the allocation and scheduling of RAN workloads. The aim is to optimize the use of GPU-based compute across the RAN environment, increasing efficiency while ensuring that AI applications and RAN functions share and co-exist on the same hardware fabric. The underlying idea is to fuse the traditionally separate domains of radio access networking and AI workloads into a unified, autonomic system where decisions about where and how to run workloads are driven by data, models, and policy rules.
This approach begins with rich telemetry from the RAN, transport, edge, and orchestration layers. Data streams are ingested into AI models that learn patterns of demand, latency constraints, service level requirements, and energy usage. The AI then informs the orchestration layer about where to place AI inference tasks, where to run RAN control functions, and how to scale resources up or down in response to traffic bursts, mobility patterns, and evolving network slicing needs. The result is a dynamic, self-optimizing platform where RAN performance and AI workloads are co-optimized in real time.
From an architectural standpoint, AI-RAN demands a tightly integrated stack that combines RAN function virtualization, AI inference engines, and a robust data fabric. The GPU servers are not merely accelerators for isolated AI workloads; they are central to a compute fabric that supports both RAN control plane and user plane tasks alongside AI services such as predictive maintenance, traffic forecasting, anomaly detection, and autonomous admission control. The architecture must also include secure, low-latency communications between the RAN functions and AI agents, ensuring that decisions are timely, auditable, and aligned with service-level commitments.
A critical aspect of AI-RAN is its deployment strategy across edge-to-core hierarchies. At the edge, latency-sensitive AI inference can be used to optimize radio resource management in radio units and distributed units. In core and central sites, more sophisticated AI models can run on more capable GPU pools to handle heavier workloads, long-horizon planning, and cross-domain optimization across multiple network domains and tenants. This hierarchical orchestration supports a spectrum of use cases—from rapid retries to proactive optimization based on predictive signals—while maintaining strict control of data governance and security.
Benefits and ROI
The AI-RAN approach promises several compelling benefits for telcos and their enterprise customers. First, resource efficiency improves as AI determines the most effective mapping of RAN functions and AI tasks to available GPU capacity, reducing waste and lowering operational costs. Second, network performance and quality of service can be raised through more informed scheduling decisions that minimize latency, prevent congestion, and improve user experiences in high-demand scenarios such as urban dense areas, stadiums, and industrial campuses. Third, AI-RAN supports more flexible and cost-effective network slicing, enabling operators to tailor performance to specific needs—elevating the potential for new revenue streams while better protecting existing service level commitments.
A related benefit is the potential to shorten time-to-value for new services. With AI-guided orchestration, telcos can accelerate the rollout of AI-enabled features (for example, predictive maintenance for network gear, pro-active capacity planning, or anomaly detection), reducing the time between concept and customer-ready service. The consolidation of RAN and AI workloads onto a common compute platform also simplifies management, potentially lowering total cost of ownership (TCO) through unified operations, shared security controls, and streamlined lifecycle management.
From a financial perspective, AI-RAN can improve the return on investment for 5G deployments by enabling more efficient use of existing assets and reducing the incremental capex needed to meet rising demand. The capability to scale AI workloads on GPU servers in tandem with RAN workloads means operators can more readily justify and realize investments in AI-powered features and services. Moreover, by enhancing network performance and reliability, AI-RAN supports higher customer satisfaction, lower churn, and stronger competitive positioning.
Implementation Considerations
Deploying AI-RAN requires careful attention to interoperability, standardization, and governance. Operators must ensure that RAN functions from multiple vendors can share a coherent data and control plane with AI workloads, and that orchestration platforms can reliably manage heterogeneous components across edge and core sites. Security is another critical dimension; as AI models access sensitive telemetry and control paths, robust authentication, provenance tracking, and model risk management are essential to safeguard against tampering and data leakage.
Interoperability with existing Open RAN ecosystems is a central consideration. Fujitsu positions its AI-RAN initiative within an Open RAN framework, emphasizing openness as a way to avoid vendor lock-in while enabling better collaboration across the ecosystem. This means adopting standardized interfaces, ensuring compatibility with common AI frameworks, and designing modular components that can be swapped or upgraded as technology advances. Data governance—particularly the collection, storage, processing, and sharing of RAN and AI data—must be explicitly defined to comply with regulatory requirements and enterprise privacy expectations.
Operationally, telcos must build or augment their analytics and AI capabilities to support AI-RAN. This includes data pipelines, feature stores for AI models, model training and validation environments, and robust monitoring for AI-augmented systems. Talent and organizational alignment are equally important: engineers, network operations personnel, and data scientists must collaborate effectively to maintain and evolve AI-RAN deployments.
The strategic implications are significant. By embedding AI into the core of RAN resource management, operators can unlock new levels of efficiency, efficiency that translates into better service quality and, crucially, more predictable and scalable returns on 5G investments. In Fujitsu’s framing, AI-RAN is not a single product but a holistic capability set that integrates the company’s R&D strengths with customer-focused implementation know-how, aligning technology possibilities with business outcomes.
Photonics, Open RAN, and Fujitsu’s Open Portfolio
Photonics: High-Capacity Data Transport and Processing
A central thread in Fujitsu’s Barcelona showcase is the use of photonics to carry vast data payloads over fibre optic networks. Photonics refers to the use of light-based technologies to move and manipulate data at extremely high speeds and low latency. In the context of 5G and AI-enabled networks, photonics provides the bandwidth backbone that allows massive data streams—from user-plane traffic to sensor data for AI analytics—to move efficiently between edge sites and central data centers. The emphasis on photonics aligns with the need to keep pace with the data deluge generated by AI-driven network optimization, real-time analytics, and the deployment of AI-based services at scale.
Fujitsu’s positioning suggests a strategy where photonics is not an isolated capability but an essential enabler of the broader AI- and Open RAN-based architecture. The ability to reliably transfer large volumes of data with minimal delay supports higher-fidelity AI inference, more frequent model updates, and more responsive orchestration decisions. In practice, this translates to improved user experiences, better support for mission-critical applications, and improved capacity management across dense urban networks where data throughput requirements are formidable.
Open RAN: An Open, Collaborative Portfolio
Open RAN remains a cornerstone of Fujitsu’s approach, with the company presenting itself as an early advocate and a strong supplier with a robust portfolio. Open RAN aims to decouple software from hardware, enabling operators to mix and match components from different vendors while maintaining interoperable interfaces. Fujitsu’s Open RAN portfolio includes software-defined control, virtualized network functions, and a suite of integration capabilities designed to work seamlessly with third-party components. The company emphasizes the ability to deliver a complete, integrated solution that respects the openness philosophy while providing the reliability and performance that operators require for commercial deployments.
The Open RAN strategy dovetails with the AI-RAN concept by enabling a flexible, multi-vendor environment where AI-augmented RAN components can coexist with other RAN elements. This openness can help operators avoid IP fragmentation and reduce procurement costs, while still enabling the sophisticated orchestration and AI-driven optimization that modern networks demand. Fujitsu’s narrative underscores the belief that openness does not come at the expense of performance or control; rather, it creates a richer ecosystem in which RAN, AI, and network management tools can collaborate more effectively.
Use Cases and Industry Impact
The combination of photonics, AI, and Open RAN yields several compelling use cases for telcos and enterprise customers. In a typical scenario, a telco can transport sensor and telemetry data from AI-accelerated RAN components with ultra-low latency using photonic links, feeding AI models that predict traffic spikes, optimize radio resource allocation, and pre-emptively adjust network slices to meet service-level agreements. In parallel, Open RAN infrastructure can be tuned through AI-driven policies to optimize energy consumption, handle dynamic capacity planning, and manage multi-vendor integration challenges without sacrificing reliability.
For enterprises adopting AI-driven transformation, the Fujitsu Open RAN portfolio supports private networks and hybrid deployments that combine public and private infrastructure with AI-enabled management. Private networks can leverage photonics-backed backhaul for large-scale data movement, while AI-RAN features help ensure that critical industrial services—such as automated manufacturing processes, predictive maintenance, or remote monitoring—receive the required reliability and performance. By aligning photonics capacity, Open RAN flexibility, and AI-driven orchestration, Fujitsu positions itself to address both the public network operator segment and the enterprise private-network market, supporting a broad spectrum of 5G-enabled innovations.
Operational and Strategic Implications
Strategically, embracing photonics and Open RAN positions Fujitsu to offer a comprehensive technology stack that can scale with operator demands. The integration of high-bandwidth fibre transport with AI-augmented RAN control enables more aggressive deployment of network slices, more responsive quality of service, and more agile deployment of new services. The openness of the platform fosters collaboration with ecosystem partners, accelerates innovation cycles, and may reduce time-to-market for new capabilities. Operationally, operators can benefit from unified lifecycle management across multi-vendor components, improved visibility into network health through AI analytics, and the ability to forecast capacity needs more accurately.
Fujitsu’s narrative also signals an emphasis on governance, security, and resilience within an Open RAN context. The alliance between photonics-enabled data transport, AI-driven network optimization, and standardized interfaces requires robust security models and clear accountability. Operators must ensure that their data governance frameworks, model risk management practices, and compliance controls align with regulatory and corporate requirements. In this landscape, Fujitsu’s integrated approach—combining research, R&D, and customer-centric deployment expertise—aims to reduce friction in the adoption of photonics-backed, AI-enhanced, Open RAN systems.
Private 5G: Enterprise Connectivity and AI Transformation
Lowering Barriers to Private Networks
A key growth theme at MWC 2025 is the maturation of Private 5G networks as a practical, scalable solution for enterprises pursuing AI-enabled transformation. Fujitsu’s private 5G offering is positioned as a way to address increasing connectivity demands in organizations that deploy AI initiatives across manufacturing floors, logistics hubs, healthcare facilities, campuses, and other mission-critical environments. The cost of private networks has been on a downward trajectory, making deployments feasible not only for large enterprises but also for medium-size and even smaller organizations that need dedicated, secure, low-latency connectivity for AI workloads.
In this context, private 5G is not simply about connectivity; it is a platform for AI-enabled automation and intelligent decision-making at the network edge. The architecture typically involves on-prem or edge cloud deployments of 5G core and radio access functions, paired with AI-enabled management and orchestration to drive automation, monitoring, and predictive analytics. The key benefit is that data remains within the enterprise perimeter or a trusted edge zone, reducing exposure to public networks and enabling more stringent latency and privacy controls. The economic argument hinges on lower total cost of ownership once deployed at scale and enabled by ongoing advances in 5G technology and AI integration.
Enterprise Use Cases and Value Creation
In practice, private 5G networks support a wide range of applications. In manufacturing, private networks enable autonomous robotic cells, real-time machine-to-machine coordination, and sophisticated defect detection and corrective actions, all powered by ultra-reliable low-latency communications (URLLC) and edge AI. In healthcare, private networks can enable remote diagnostics with near-instantaneous data exchange, real-time telemedicine, and secure patient data flows that are compliant with stringent privacy standards. Logistics and warehousing benefit from precise asset tracking, automated inventory management, and reliable, on-site AI analytics that optimize operations.
Fujitsu emphasizes that the value of private 5G comes not only from the network itself but also from its integration with AI capabilities across the enterprise. The ability to deploy edge AI workloads in conjunction with a secure private 5G infrastructure creates opportunities for accelerated decision-making, optimized processes, and new service offerings that can justify CAPEX and OPEX investments. As the cost curve continues to improve, private 5G is increasingly accessible to a broader set of organizations, making it a critical component of the broader digital transformation journey.
Ecosystem, Security, and Deployment Patterns
A practical private-5G strategy requires careful attention to security, interoperability, and deployment models. Fujitsu’s approach encompasses end-to-end solutions that consider the spectrum from radio access to core and edge, with AI-enabled management and orchestration to optimize performance and reliability. Operators and enterprises must coordinate across multiple stakeholders, including network equipment vendors, system integrators, cloud providers, and software developers, to ensure smooth integration of AI applications with private networks. Security measures must cover device onboarding, network segmentation, data governance, encryption, and continuous monitoring to detect and mitigate threats in real time.
Deployment patterns for private 5G span a spectrum from on-site campus networks to private networks hosted in carrier-enabled edge-to-core ecosystems. In some scenarios, hybrid deployments that combine on-prem components with secure, private cloud resources at the edge enable organizations to balance performance, cost, and scalability. Fujitsu’s private 5G solutions are designed to be adaptable to these patterns, enabling operators and enterprises to tailor configurations to their specific needs while maintaining a coherent management and security framework.
Strategic Rationale for Telcos and Technology Providers
For telecom operators and technology suppliers, private 5G represents a strategic opportunity to monetize AI-driven transformation. The capability to offer managed services and end-to-end solutions—including private networks, edge AI, and integrated orchestration—opens avenues for revenue beyond traditional connectivity. In this context, Fujitsu positions its private 5G portfolio as a key enabler for the broader AI-augmented network strategy, aligning enterprise needs with telco capabilities to deliver measurable ROI and a clearer path to monetization of 5G investments.
5G Standalone: Industry Transformation and Automation
5G Standalone and Its Distinctive Advantages
Fujitsu’s discussions at MWC 2025 underscore that 5G Standalone (SA) represents a meaningful leap beyond Non-Standalone (NSA) deployments. SA brings a more efficient, flexible, and capable network architecture with features such as enhanced mobile broadband, ultra-reliable low-latency communications, and network slicing that can be tuned to different sector-specific requirements. The return on investment for telcos and their customers is tied to the ability of SA to unlock capabilities that were previously unattainable or prohibitively expensive on 4G-based frameworks.
From a technology perspective, SA enables lower latency, higher throughput, and more deterministic performance, unlocking new workflows in manufacturing, logistics, healthcare, and beyond. It also facilitates more sophisticated automation and orchestration strategies, as the control plane becomes more capable and responsive to real-time conditions. The impact extends to industries that require precise timing, synchronized operations, and high levels of automation, all of which can be achieved more reliably on SA networks.
Industry Transformation Across Sectors
The transformative potential of SA is not limited to telecoms; it extends to a broad set of industries. In manufacturing, SA can support intelligent factories where robotic systems, sensors, and AI agents coordinate in real time to optimize production lines, reduce waste, and improve quality. In healthcare, SA enables remote monitoring, telepresence, and real-time data exchange with clinical systems, creating opportunities for improved patient outcomes and efficiencies. Across agriculture, energy, and smart cities, the combination of SA, AI, and data analytics drives new levels of automation, predictive analytics, and decision support.
The common thread across these sectors is the need for reliable, high-capacity, and low-latency networks that can carry AI-driven workloads at the edge and in the core. SA provides the foundation for these capabilities, enabling more sophisticated network services, better service quality, and the ability to deploy new, revenue-generating AI-enabled applications with confidence.
Operational and Management Implications
Realizing the benefits of SA requires robust network management and operation practices. Operators must invest in orchestration platforms capable of end-to-end management across the network, the edge, and the cloud, with AI-enhanced analytics to drive continuous optimization. This includes proactive capacity planning, service assurance, and autonomous remediation of network issues. The integration of AI into these management workflows is essential to achieving scalable, repeatable outcomes and to lowering the total cost of ownership associated with complex SA deployments.
From Fujitsu’s perspective, the combination of SA with AI-RAN and Open RAN fosters a cohesive, end-to-end capability stack. This stack combines the computational power of AI-enabled RAN coordination with the openness and flexibility of multi-vendor ecosystems and the high-capacity transport enabled by photonics. The net result is a platform that not only delivers superior network performance but also creates opportunities for operators to offer new, AI-backed services that differentiate them in a highly competitive market.
Market Dynamics: ROI, Telcos, and New Revenue Streams
ROI as a Central Guiding Principle
A recurring theme in Fujitsu’s MWC narrative is the central role of return on investment when evaluating 5G and AI-driven technologies. Operators are seeking tangible, measurable improvements in ROI from their network investments, particularly as they navigate the transition to AI-enabled networks, Open RAN architectures, and private network deployments. Fujitsu positions its portfolio as a way to maximize ROI by combining resource efficiency with new capabilities such as AI-enabled automation, private-network monetization, and value-added services built on 5G SA and AI.
The path to ROI is not purely technical. It involves operational discipline—effective governance, cost control, and performance measurement—as well as strategic alignment with enterprise customer needs. Operators must balance necessary capital investments with the promise of long-term operational savings and new revenue streams. In this framework, AI-RAN, photonics-enabled transport, Open RAN openness, private networks, and SA capabilities all contribute to a more compelling business case when properly integrated and managed.
European Landscape and Industry Consolidation
The European telco landscape presents both opportunities and challenges. There is an ongoing tension between the desire to maintain competitive, locally tailored networks and the potential benefits of greater scale through consolidation and cross-border collaboration. There is discussion about the number of telcos in some countries and how consolidation or restructuring could yield more efficient operations, accelerate technology deployments, and improve ROI in a competitive global market. In this context, Fujitsu’s emphasis on integrated, open, and scalable platforms could help operators navigate a more concentrated future by enabling flexible consolidation strategies without sacrificing interoperability or service quality.
Growth Opportunities for Vendors and Operators
For technology providers like Fujitsu, the market presents multiple growth avenues. There is demand for end-to-end solutions that combine AI, Open RAN, photonics transport, Private 5G, and network management in a cohesive package. Operators are seeking partners who can deliver integrated capabilities with proven performance, robust security, and clear ROI. The opportunities extend beyond the network core to new services and business models, including managed services, network-as-a-service offerings, and AI-enabled operations that help enterprises realize the value of 5G and automation in their own operations.
Vendors can differentiate themselves through the depth of their integration capabilities, the breadth of their ecosystem partnerships, and the ability to deliver practical, scalable deployments that align with operator and enterprise goals. Fujitsu’s strategy—rooted in AI, openness, and enterprise-grade execution—resonates with this market demand by offering not just components but a complete, deployable solution that bridges research with customer outcomes.
Deployment Realities: AI Adoption, Disruption, and the Path Forward
AI Adoption in the Enterprise and Telco Ecosystem
Despite the proliferation of AI technologies at major industry events, the real-world deployment of AI across telcos and enterprises remains uneven. While AI is a prominent topic at MWC and similar venues, there is a sense that truly disruptive AI-enabled capabilities are still emerging rather than widely deployed. Telcos face a range of challenges—from integration complexity and legacy systems to data governance, security, and workforce readiness—that can slow down AI adoption. At the same time, AI-enabled network optimization, AI-assisted management, and AI-based automation hold significant promise for improving network performance, efficiency, and service innovation.
The path forward involves building repeatable, scalable deployment patterns that can be adapted to different operator contexts. It also requires a thoughtful approach to change management, including upskilling staff, establishing clear governance for data and models, and creating measurable metrics to track ROI and performance improvements. Operators and vendors alike need to pursue pilots, proofs of concept, and phased rollouts that demonstrate tangible business value and build confidence for broader adoption.
Disruption Versus Incremental Improvement
The industry recognizes that AI and 5G are transformative, but achieving disruptive change requires more than technology. It demands new operating models, cross-functional collaboration, and a willingness to rethink how networks are designed, managed, and monetized. Fujitsu’s framing suggests that AI-RAN and related capabilities can catalyze this transformation by enabling more intelligent, autonomous network operations, faster service delivery, and more responsive control over resource allocation. For telcos, this implies not only new technical capabilities but also the opportunity to reshape business models toward value-driven outcomes rather than purely capital-intensive infrastructure.
The Roadmap for Operators and Partners
Realizing the benefits of Fujitsu’s AI-led 5G strategy will require a well-planned roadmap that aligns technology milestones with business objectives. Operators should identify the most compelling initial use cases for AI-RAN and Open RAN that deliver clear ROI and user impact. They should plan data governance and security milestones in parallel with architectural milestones to ensure compliance and risk management. Partners and ecosystem players can contribute by delivering interoperable components, co-developing AI models for RAN optimization, and offering end-to-end services that reduce integration risk for operators.
A successful roadmap also involves stakeholder education and market communication. Operators must articulate the value proposition of AI-enhanced networks to enterprise customers, highlighting how private networks, AI-powered automation, and 5G SA capabilities translate into tangible business benefits. For technology providers like Fujitsu, continuous investment in R&D, meaningful engagement with operators, and a focus on practical deployment patterns will be essential to sustaining momentum and achieving long-term growth.
Conclusion
Fujitsu’s MWC 2025 presentation underscored a cohesive, AI-centric vision for 5G that aims to redefine how telcos realize value from their networks. Central to this vision is AI-RAN, a strategic integration of radiowave infrastructure with AI workloads that uses GPU resources more efficiently to deliver smarter, more responsive network behavior. The approach blends two key technologies—Open RAN and photonics—to build an open, high-capacity, and flexible platform that can scale across edge and core, across public and private deployments, and across multiple industries.
Private 5G networks emerge as a critical component of this strategy, lowering barriers to enterprise AI adoption by delivering secure, low-latency connectivity in a controlled environment. Meanwhile, 5G Standalone offers the performance and control required to unlock advanced automation and sector-specific capabilities that transform industries such as manufacturing and healthcare. In parallel, Fujitsu’s Open RAN portfolio provides an openness-driven approach that supports multi-vendor interoperability while maintaining reliability and execution discipline.
The broader market narrative reinforces the importance of return on investment, network management, and the creation of new revenue streams for telcos. Operators must balance the costs and benefits of AI-enabled networks with strategic considerations about European market dynamics, competition, and regulatory contexts. For technology providers, the opportunity lies in delivering end-to-end, integrated solutions that combine AI, RAN, and high-capacity transport while reducing deployment friction and maximizing enterprise value.
As telcos and enterprises navigate the next wave of 5G-enabled transformation, Fujitsu positions itself as a partner able to bridge research with real-world deployment, linking AI capabilities, Open RAN openness, and photonics-powered data movement into a unified platform. The result is a compelling pathway toward faster time-to-value, smarter networks, and broader, more resilient digital ecosystems that can support the AI-driven future of connectivity.