Riyadh: The rapid expansion of AI workloads is driving a fundamental shift in data center network infrastructure, with global data center experts forecasting a dramatic rise in interconnect bandwidth demand over the next five years, according to a study commissioned by Ciena (NYSE:CIEN).
Conducted in partnership with Censuswide, the survey gathered insights from over 1,300 data center decision-makers across 13 countries. More than half (53%) of respondents identified AI workloads as the biggest driver of demand for data center interconnect (DCI) infrastructure over the next 2-3 years, outpacing cloud computing (51%) and big data analytics (44%).
To accommodate the growing AI-driven demand, 43% of new data center facilities are expected to be dedicated to AI workloads. With AI model training and inference requiring significant data movement, experts predict a substantial surge in bandwidth requirements. Additionally, 87% of participants indicated the need for fiber optic capacity of 800 Gb/s or higher per wavelength to support these expanding requirements.
Sustainable AI-Driven Networks Survey respondents highlighted the increasing role of pluggable optics in meeting bandwidth demands while addressing power and space limitations. A striking 98% of data center experts consider pluggable optics essential for reducing power consumption and minimizing network infrastructure footprints.
Distributed Computing and AI Training As AI compute requirements escalate, training for Large Language Models (LLMs) is expected to be increasingly distributed across multiple AI data centers. According to the survey, 81% of respondents anticipate LLM training occurring across interconnected data center facilities.
When asked about the primary factors influencing AI inference deployment, respondents ranked the following priorities: AI resource utilization over time (63%), Reducing latency by placing inference compute closer to users at the edge (56%), Data sovereignty requirements (54%) and Strategic locations for key customers (54%)
Rather than deploying dark fiber, a majority (67%) of respondents expect to utilize Managed Optical Fiber Networks (MOFN), which leverage carrier-operated high-capacity networks for long-haul data center connectivity.
"The AI revolution is not just about computers, it’s about connectivity," said Jürgen Hatheier, Chief Technology Officer, International, Ciena. "Without a strong network foundation, AI’s full potential cannot be realized. Operators must ensure their DCI infrastructure is prepared for a future where AI-driven traffic dominates."
