Mobile World Congress 2026 delivered a clear signal: the infrastructure conversation in telecommunications is evolving from speculative theory to operational reality. Operators are no longer discussing these trends only at a conceptual level. Many are actively evaluating how their infrastructure platforms must evolve to support AI and modern workloads, stricter regulatory requirements, and changing economics across compute, memory, and storage.  Across sessions and in-depth conversations with operators, partners, analysts, and industry leaders, four priorities consistently surfaced throughout the week:

Infrastructure economics and Total Cost of Ownership (TCO) are becoming central factors in operator planning. Several operators described memory and storage as growing contributors to cost, particularly as AI workloads introduce new resource demands.

Telco infrastructure discussions are increasingly centered on leveraging software capabilities to boost hardware efficiency and achieve long-term economic sustainability. Operators are prioritizing software-centric architectures that allow for the introduction of new services without relying solely on hardware refreshes. By running functions on general-purpose hardware and using virtualization to scale resources dynamically, operators can improve agility and reduce capital constraints.

Key software-defined capabilities include:

• Advanced NVMe Memory Tiering: Extending system memory by using high-performance NVMe devices as a secondary layer. This allows operators to increase capacity while managing DRAM costs.
• Global Deduplication: Reducing storage consumption by eliminating redundant data across distributed infrastructure environments rather than within individual clusters.
• Data-Driven Automation: Implementing intelligent automation to reduce complexity. Efficient automation requires a foundation of high-quality telemetry and unified data. If automation is based on inaccurate data, it can lead to operational failures such as loops or service degradation.

By focusing on software flexibility and unified data domains, operators can achieve the performance required for modern workloads while maintaining control over operational and infrastructure costs.

Broadcom’s Ram Velaga, President of Infrastructure Software, identified three strategic shifts in telco infrastructure at MWC 2026: the rise of sovereign AI, the move to private cloud, and deployment challenges (power, space, and silicon). He announced Broadcom’s new Telco Cloud Platform 9, which unifies VMs and containers on the same platform while cutting physical infrastructure costs.

Watch the full discussion

At MWC 2026, sovereignty shifted from an abstract regulatory concept to a substantial commercial value driver. In an environment of geopolitical and digital uncertainty, sovereign AI and sovereign cloud have transitioned into critical requirements. This strategic pivot redefines sovereignty beyond simple data residency. Increasingly, the focus is on the ownership and operation of the digital infrastructure itself. Organizations now seek assurance that critical services are hosted on platforms managed by trusted, local personnel within their own jurisdictions.

Telecommunications operators are uniquely positioned to address this market by leveraging their core assets:This strategic pivot redefines sovereignty beyond simple data residency. Increasingly, the focus is on the ownership and operation of the digital infrastructure itself. Organizations now seek assurance that critical services are hosted on platforms managed by trusted, local personnel within their own jurisdictions.

Telecommunications operators are uniquely positioned to address this market by leveraging their core assets:

• Infrastructure: Repositioning carrier-owned data centers as secure platforms for regulated cloud and AI workloads.
• Regulated networks: Utilizing regulated networks and licensed spectrum to build a sovereign tech stack that differs from global, non-regional providers.
• Localized Governance: Hosting regulated data and AI services close to the point of consumption to adopt advanced capabilities without losing operational control.

By aligning infrastructure with these sovereignty requirements, operators can move up the value chain shifting from standalone bandwidth providers to trusted platform partners for the most sensitive digital workloads.

Paul Turner, Chief Product Officer for the VMware Cloud Foundation Division at Broadcom, discusses the strategic demands on telco operators. In an interview, he explains how next-generation VMware Telco Cloud Platform gives operators architectural control, crucial for meeting sovereign requirements. He details what operators need to manage distributed infrastructure, navigate regulations, and prepare for AI-native workloads.

Watch the full discussion

The discussion around AI in telecommunications has moved beyond proof of concepts. Operators are increasingly assessing how infrastructure can support AI workloads alongside traditional network services.

A recurring theme is that moving from AI pilots to operational deployments introduces practical challenges. The focus is shifting toward AI Native, where machine learning is embedded directly into the core architecture rather than treated as an external tool. Building these intelligence-driven networks requires breaking down traditional silos between data, operations, and network domains. Only with unified visibility can networks deliver the scale required to support an AI-driven ecosystem.

The goal is to move beyond simple decision-support tools and toward a foundation for autonomous operations. This includes enabling predictive fault detection, intelligent resource allocation, and automated network optimization. Instead of over-engineering capacity based on speculative demand, many operators are starting with specific use cases and defined user requirements. This approach allows infrastructure investments to scale with real workloads rather than theoretical projections.

A primary challenge is the diversity of AI workloads, which require different placement strategies:

• Training workloads: Typically centralized in large-scale environments capable of managing massive data sets and high-density compute power.
• Inference workloads: Often distributed across the network to meet low-latency needs for specific end-user applications in sectors like healthcare, manufacturing, and automotive.

Operators are also evaluating how to run both network functions and emerging AI workloads on unified infrastructure platforms. Operating multiple infrastructure stacks increases both operational complexity and capital requirements.

Anupama Mahabhashyam of Broadcom, discusses the operational realities of building private and sovereign AI clouds. She explains how operators can transform regulated infrastructure into a durable AI advantage through rigorous model governance and data readiness with VMware Private AI as a Service. 

Watch the full discussion here 

4: Carrier-Grade Cloud Native Operations

A significant shift observed at MWC 2026 is growing operator confidence in Kubernetes for production network function deployment. Early telco Kubernetes initiatives faced operational challenges around upgrade complexity, support lifecycles, and service continuity during platform maintenance. While those concerns remain relevant, platform capabilities and operational practices are reaching higher levels of maturity.

Cloud-native architecture has evolved from a method for rapid compute acquisition into a complete platform transformation. In this environment, Kubernetes acts as the standard for desired-state automation. It determines how software scales, heals, and interacts with the underlying compute, storage, and networking fabric. A primary trend is the rising demand for a consistent cloud experience on-premises, specifically as networks transition from Virtual Network Functions (VNFs) to Cloud-Native Network Functions (CNFs). 

Dilpreet Bindra of Broadcom explores the transition of Kubernetes into a mainstream operational standard. He details how the integration of automation and multi-tenant operations creates a more unified interface for both enterprises and service providers.

Watch the full discussion

Partner Ecosystem

VMware Telco Cloud Platform relies on a robust partner ecosystem to ensure that operators have access to validated, high-performance solutions. Collaborations with partners like Network Equipment Providers (NEPs) and silicon vendors, are essential for delivering the interoperability and performance required for modern virtualized and containerized environments. These partnerships allow operators to deploy proven network functions on a consistent infrastructure layer, reducing the risk and complexity of multi-vendor environments.

Recent highlights from our ecosystem include:

• Nokia and Broadcom Collaboration: Gordon Milliken of Nokia and Padma Sudarsan of Broadcom discuss how Broadcom and Nokia are working together to support the shift to cloud-native architectures. Their joint efforts address operational complexity, ensure comprehensive validation, and enable operators to run both VM-based and cloud-native solutions during the transition. Watch the partnership discussion
• Expanding Choice with Canonical: We continue to collaborate with Canonical to broaden the choice for open-source infrastructure and OS layers within the telco stack. Read the partnership details 

Establishing these deep technical integrations ensures that the platform remains a versatile and reliable foundation for the diverse requirements of global operators.

In a discussion with Evan Kirstel, Odded Solomon and Madhup Gulati of Broadcom delve into the ongoing transformation of the telco sector. Their detailed analysis outlines a practical strategy for future developments, focusing on the shift towards software-driven networks and utilizing upcoming platform enhancements to better monetize 5G and AI assets.

Watch the discussion here

To navigate the shifts observed at MWC 2026, operators must transition from exploration to the scalable deployment of modern infrastructure. Capitalizing on these emerging trends requires prioritizing the following strategic actions:

• Operationalize the AI Infrastructure Roadmap: Focus on transitioning from isolated AI experiments to a unified production environment. Prioritize high-impact workflows that allow for the automated management of complex network tasks, reducing operational layers and aligning infrastructure capacity with actual workload demands.
• Commercialize Sovereign Tech Stacks: Position localized data residency and secure governance as high-value. By offering infrastructure that is managed and hosted locally, operators can provide a differentiated competitive advantage for government and regulated industry clients prioritizing digital autonomy.
• Evolve Toward a Unified Platform Model: Long-term growth will depend on the ability to support diverse workloads, including both VNFs and CNFs network functions and AI workloads, on a single, economically efficient platform. Success requires moving beyond traditional bandwidth delivery toward engineering the intelligent and sovereign platforms that will define the next decade of telecommunications.

Learn More

For more information about VMware Telco Cloud Platform and Broadcom’s approach to telecommunications infrastructure, visit the Broadcom Telco Cloud resources or contact your Broadcom representative.