Leading Tier-1 network operators are advancing toward AI-native networks based on their existing cloud-native foundations, according to a report by Omdia. This transformation goes far beyond an architectural upgrade; it marks a fundamental shift in how networks are operated, managed, and monetized.

Omdia predicts that by 2031, AI functions will account for 12% of network application spending, driving global telecom operators' investment in cloud and AI infrastructure to $22.9 billion. Achieving AI-native operations requires comprehensive transformation across multiple layers: network applications must evolve to embed intelligence; cloud platforms need to extend beyond Container-as-a-Service (CaaS) to encompass Machine Learning Operations (MLOps) and full model lifecycle management; management and orchestration systems must shift from rule-based automation to agentic reasoning; and integration services must also adapt to support multi-agent workflows and closed-loop autonomous operations.

For operators navigating this change, understanding how AI-native architecture will reshape existing cloud-native technology stacks is critical to maintaining competitive advantage, meeting regulatory compliance requirements, and opening new revenue streams. The report systematically examines the key architectural changes that AI-native networks will bring, their impact on existing cloud-native technology stacks, and the priority investment areas in the era of Autonomous Networks (AN).

As operators shift from application-centric cloud platforms to model-and-data-centric AI-native infrastructure, the communications industry is undergoing a fundamental architectural transformation, according to Omdia's perspective. Omdia forecasts that telecom operators' spending on cloud-native and AI-native networks will grow at a 9% compound annual growth rate (CAGR) from $14.9 billion in 2026 to $22.9 billion in 2031.

This growth reflects three parallel investment drivers: continued modernization and migration of traditional network functions to cloud-native architectures; integration of AI into network operations tools; and preparation for 6G infrastructure deployment in the latter part of the forecast period. The architectural shift from cloud-native to AI-native networks is substantial. Traditional cloud-native platforms prioritize container orchestration and application lifecycle management, while AI-native platforms must handle model training pipelines, inference serving, feature stores, vector databases, and continuous model retraining. These requirements impose fundamentally new demands on resource scheduling, data management, and governance frameworks.

The operational impact extends beyond infrastructure procurement to management systems and service delivery models. As operators transition from dedicated virtualization platforms to shared multi-tenant cloud environments, security and regulatory compliance frameworks such as the Network and Information Systems Directive 2 (NIS2) are driving architectural requirements around isolation, segmentation, zero-trust implementation, and supply chain security assurance.

The CaaS layer is undergoing the most significant evolution: MLOps is emerging as a distinct spending category alongside traditional cloud-native network function (CNF) and virtual network function (VNF) hosting, enabling operators to manage the full AI model lifecycle, automate retraining pipelines, and deliver production-grade inference services. Management and orchestration systems are embedding closed-loop autonomous capabilities: on one hand, AI enhances fault detection, root cause analysis, and decision recommendations; on the other hand, cloud-native frameworks such as GitOps handle the execution of automated operations.

This operational transformation is also reshaping demand for professional services. Omdia expects that through 2031, systems integration and professional services spending will grow at a 13% CAGR, with operators shifting focus from traditional cloud migration projects to data platform implementation, observability frameworks, AI model customization, and automated pipeline construction.

Key Takeaways

Driven by AI and cloud-native transformation, spending on cloud-native and AI-native networks is projected to grow to nearly $23 billion by 2031. Growth remains cyclical, and we expect acceleration after 2029 as operators increase investment in AI-enabled network capabilities and cloud-native architectures to support 6G deployment.

Application software is becoming the dominant spending category. Application software (VNF, CNF, AI-enabled network functions) spending will grow from $6.2 billion in 2026 to $11.1 billion in 2031, at a 12% CAGR, making it the fastest-growing major segment. Its share of total spending will rise from 42% in 2026 to 48% in 2031, indicating that telecom operators are shifting value from infrastructure toward software-defined, intelligent network functions.

Hardware spending is expected to remain elevated in 2026 and 2027. Cloud and AI infrastructure spending will grow 15% year-over-year in 2026, while application software will grow 9% year-over-year. Multiple factors, including tight supply of DRAM, high-bandwidth memory (HBM), and CPUs, are affecting server economics and procurement timing.

CNF is overtaking VNF. Telecom operators' spending on CNF will grow from $2.7 billion in 2026 to $7.2 billion in 2031, at a 22% CAGR, while VNF revenue remains essentially flat. Operators are moving from virtualization to fully cloud-native network architectures. By 2031, CNF spending will be nearly twice that of VNF spending.

AI-related spending, while starting from a small base, is growing rapidly. AI infrastructure spending for communications applications will increase from $28 million in 2026 to $742 million in 2031, at a CAGR of 92%. AI capabilities embedded in network functions are growing even faster, from $11 million in 2026 to $1.3 billion in 2031. Over the same period, MLOps/large language model (LLM) operations middleware spending will expand from $29 million to $234 million, highlighting the rise of the AI software stack in communications networks.

Professional services are becoming one of the fastest-growing segments. To operationalize cloud-native and AI-enabled networks, telecom operators increasingly need professional capabilities in systems integration, migration, and AI transformation.

Recommendations

Telecom operators must invest in cloud-native operating models as the foundation for autonomous operations. Without cloud-native operating models, AI agents cannot safely execute remediation actions at scale. Manual approval processes and imperative scripts create bottlenecks that prevent autonomous operations from delivering their promised efficiency gains. Cloud-native provides the programmable foundation that makes autonomous execution both feasible and safe.

Operators must prioritize building unified telemetry platforms using open-source software. A unified telemetry platform enables multi-signal anomaly detection, causal inference, and context-aware reasoning—capabilities that distinguish effective AIOps from simple alerting systems. Without this foundation, operators will continue to face alert fatigue, slow root cause identification, and fragmented operational visibility, making it difficult for AI to create meaningful value.