Executive Summary
- AI is set to become a major revenue driver for service providers, with 90% expecting high-capacity AI network services to drive growth over the next three to five years.
- But, urgent network investment and upgrade is needed to deliver the capacity, reliability and performance required by AI-driven services.
- AI is creating new opportunities beyond traditional connectivity, including data centre interconnect, GPU-as-a-Service, edge computing and multi-cloud services.
A recent Ciena survey of service providers (conducted in collaboration with Censuswide) questioned more than 1,200 telecom, wholesale, and regional service provider experts across 12 countries and found that service providers see multiple emerging opportunities to expand AI-related revenue streams.
The survey underscores the revenue potential of AI-driven networking. 90% of respondents expect high-capacity AI network services – connectivity to AI infrastructure required by enterprises, hyperscalers, and neoscalers – to drive revenue growth over the next three to five years, with 56% citing them as their primary source of net-new revenue.
Confidence is equally strong in managed optical fiber network (MOFN) services, with 96% expecting them to generate revenue from connecting distributed AI compute clusters over the next three years, and 51% identifying MOFN as their primary vehicle for delivering data center interconnect (DCI) services.
Strong sense of urgency
However, as service providers pursue new AI-driven revenue opportunities, the majority (88%) of survey respondents had a strong sense of urgency on the need for optical network upgrades to support the demands of premium enterprise AI service level agreements (SLAs). Additionally:
- Nearly half (48%) believe those upgrades are critical, and 39% said they are needed within the next 12-18 months.
- Only 11% believe routine upgrades will be sufficient to meet the demand.
- Advanced network automation, including agentic AI, is seen as essential (96% surveyed agree) to capitalise on AI-driven opportunities.
AI Inference
Nearly half (49%) of those surveyed believe growth in AI inference data centers will create new revenue opportunities by increasing demand for data center interconnect (DCI) services.
GPU-as-a-Service
Service providers see significant opportunities to capture GPU-as-a-service revenue, with 94% viewing strategic partnerships with cloud and neocloud providers as essential to success. Among those surveyed, 44% plan to adopt revenue-sharing models as their primary go-to-market strategy.
Connecting AI in day-to-day life
The growing use of AI in everyday life is presenting additional revenue opportunities for service providers. Almost all respondents (95%) believe immersive and AI-driven entertainment will contribute to new revenue streams in the next three years. Nearly one-third (29%) named AI wearables and personal devices, including smart glasses and health trackers, as the leading consumer hardware category expected to drive premium service revenue.
The network edge
39% of service providers surveyed believe services hosted at the network edge, such as edge compute, GPUaaS and localized AI inference, will drive more than 20% of enterprise revenue over the next three years, while 88% expect it to drive at least 10%.
Physical AI
Nearly two-thirds (60%) of respondents expect physical AI ecosystems, including industrial robotics and remote-control systems, to account for more than 15% of their total enterprise AI revenue within five years. Beyond revenue opportunities, the survey highlights how AI is reshaping service provider priorities, from network performance and cloud connectivity to security and new service delivery models. Additional findings include:
The race for reliability, not just bandwidth
AI workloads are increasingly critical to business processes and the networks supporting them must be capable of delivering high-performance connectivity consistently to keep applications running satisfactorily. 35% of service providers identify network consistency – including guaranteed performance stability and low jitter – as the top opportunity for monetization through premium service offerings.
The rise of scale-across AI infrastructure
As power constraints limit data center expansion on a single campus, hyperscalers are scaling compute capacity across distributed locations. These facilities need to be tightly integrated requiring ultra-high-speed synchronous communications. In the case of distributed synchronous training of large frontier models, these connections can reach tens of petabits-per-second, requiring up to hundreds of fiber pairs lit in parallel at full capacity. 95%[6] of the respondents see hyperscaler scale-across demand as a substantial contributor for wholesale revenue growth.
Revenue drivers for multi-cloud enterprise environments
Service providers are shifting toward dynamic connectivity models as their main revenue drivers. Multi-cloud connectivity management (56%) and consumption-based bandwidth services (50%) rank highest, outpacing traditional fixed-capacity services (38%). Additionally, 47% expect enterprises moving large datasets between cloud environments for AI processing to drive demand for high-capacity, flexible cloud connectivity.
Quantum-safe encryption is critical
Secure networking remains top of mind with quantum-safe technology emerging as a key focus for service providers preparing their networks for AI-driven demand. More than half of respondents (51%) stated that they have already launched or expect to launch commercial quantum-safe encryption services, including QKD-protected links and/or PQC-ready managed security, within 12 months. Another 48%[7] are in the early stages of development or evaluation, underscoring the importance of preparing for quantum computer threats. Additionally, ‘advanced physical and cyber security’ capabilities were cited as the leading commercial driver of sovereign network infrastructure.
