Reimagining Network as a Service (NaaS) for the AI era

With the seemingly insatiable demand for AI and thousands of miles of fiber at their disposal, communication service providers (CSPs) are seeing a renewed opportunity to monetize their network. However, unlocking this potential requires a new way of thinking.

Global AI infrastructure investment is projected to approach trillions of dollars across compute, data centers, power, and networking. Within this broader investment cycle, demand is accelerating for high-performance connectivity linking AI data centers, hyperscale clouds, enterprise environments, and edge locations. Leading CSPs are evolving into AI infrastructure partners, securing billions of dollars in AI networking contracts to connect distributed AI resources and support enterprise AI applications at scale.

As AI training continues to advance and foundation models mature, investment is increasingly shifting toward AI inferencing, where trained models generate predictions and responses from new data. While inference will occur across hyperscale, enterprise, and edge environments, latency-sensitive applications are driving greater demand for distributed inference closer to users and devices. This creates a significant opportunity for CSPs, whose distributed networks, edge infrastructure, and proximity to enterprise customers position them to support low-latency AI workloads at scale.

This shift places transport networks, wavelength services, and on-demand connectivity at the center of AI infrastructure. AI training, inferencing, and data movement generate unprecedented east-west traffic across distributed environments, creating a significant opportunity for CSPs to monetize their fiber assets with high-value enterprise connectivity services. AT&T, for example, plans to expand its fiber footprint to 60 million premises by the end of the decade to fuel enterprise growth.

To fully harness the opportunity, CSPs will need more than expanded network reach. Connectivity services must be provisioned, modified, and managed on-demand to support dynamic AI applications and rapidly changing traffic patterns.

As AvidThink notes in its Network as a Service (NaaS) report, CSPs are increasingly turning to NaaS to meet these requirements. The momentum is already evident: Lumen reported 25% growth in active NaaS customers in Q1 2026, with nearly one-quarter representing new customers. Let's explore how NaaS is helping CSPs capture new revenue opportunities in the AI era.

NaaS as an intelligent automation layer

NaaS has traditionally been viewed as a monetization model that enables CSPs to deliver on-demand connectivity through software-driven automation. In the AI era, however, the model is evolving into an intelligent automation layer that unifies data, workflows, AI, and network operations at scale.

Modern NaaS abstracts network complexity by decoupling service orchestration from the underlying network infrastructure, enabling AI-driven automation across multi-vendor, multi-domain environments. Through open APIs, it eliminates operational silos and proprietary integrations, allowing CSPs to deliver cloud-like, on-demand services while creating a scalable foundation for autonomous networks.

NaaS in action: Reducing service activation times by 80%

Forward-looking telecom operators such as Orange Business and Telefonica Germany are deploying NaaS to capture new enterprise AI opportunities. Blue Planet supports this transformation by providing the intelligent automation layer that unifies inventory, orchestration, assurance, and optimization on a cloud-native, AI-driven platform, across domains and vendors.

Figure 1: Blue Planet’s NaaS framework

Blue Planet customers are already realizing measurable business value, including:

• 3-5x increase in order processing volumes through automated, AI-assisted order management
• 80% reduction in service activation times with intent-based orchestration
• 95% reduction in capacity reporting time through accurate, unified inventory
• 80% reduction in 5G slice activation costs by reducing activation times from months to hours
• 85% faster service restoration through closed-loop automation

One of the most common ways Blue Planet supports NaaS offerings is by enabling CSP customers to create new enterprise high-speed services via a simple user interface. Let’s walk through the steps:

1. Service request initiated via AI-assisted customer portal: Enterprise customers request a new service via the self-service portal. This portal interfaces with Blue Planet Service Order Management (SOM) via standard TMF APIs, which, together with the Blue AI Assistant, qualifies the service in natural language.

2. Feasibility check: The service qualification automatically validates the enterprise services requirements against available network capacity, utilizing accurate and stateful network resource and service topology data in Blue Planet Inventory (BPI).

3. Dynamic planning and activation: Once the network capacity is confirmed in BPI, service scheduling and activation occur within Blue Planet Multi-Domain Service Orchestration (MDSO). It translates the user intent into service activation, automating end-to-end service provisioning across multiple domains and network layers. If the requirements of the workload change, Blue Planet can initiate the necessary adjustments.

4. Automated assurance: Once activated, service details are shared with Blue Planet Unified Assurance and Analytics (UAA), which begins to immediately monitor the performance and faults of the devices and network supporting this service. If performance degradations are detected, UAA can send alerts, request trouble tickets, or initiate closed-loop responses by interfacing with MDSO, BPI, and third-party service management platforms. This empowers CSPs to meet enterprise SLAs by taking proactive steps before problems occur.

5. Agentic AI: Underscoring each of these steps is the use of AI agents with Blue Planet AI Studio. AI Studio enables intent-driven automation by orchestrating AI agents across OSS workflows, helping CSPs build the intelligent, software-driven NaaS foundation needed to support rapidly growing AI workloads.

6. Continuous service optimization: As enterprise application demands evolve, AI agents continuously assess service performance, capacity, and SLA compliance, recommending or automatically implementing network adjustments through closed-loop orchestration. This enables CSPs to deliver adaptive NaaS experiences without manual intervention.

Figure 2: Enterprise ethernet service request

By implementing Blue Planet’s intelligent automation solution, CSPs can offer a rich NaaS solution with a simple customer interface. As user requirements and workloads change, the solution provides flexibility to adapt without creating unnecessary complexity or long lead times. This agility is exactly what is needed as the demand for AI explodes, and enterprises expect high-speed access to AI data and inferencing centers within hours, not months.

The takeaway

The new NaaS opportunity extends well beyond faster service activation. AI-driven NaaS enables CSPs to transform their networks into programmable, revenue-generating platforms that deliver the cloud-like experiences enterprises now expect. Those that modernize their OSS today will be best positioned to capitalize on the AI infrastructure economy tomorrow.