AI to dominate data centre demand, new survey reveals

A new global survey commissioned by Ciena has revealed significant expected increases in data centre interconnect (DCI) bandwidth demand driven by artificial intelligence (AI) workloads over the next five years.

The study involved 1,300 data centre decision makers across 13 countries and highlighted that AI workloads are expected to have the most significant impact on data centre infrastructure, with 53% of respondents indicating AI will surpass cloud computing and big data analytics within the next two to three years. Consequently, 43% of new data centre facilities are expected to focus primarily on AI workloads.

"AI workloads are reshaping the entire data centre landscape, from infrastructure builds to bandwidth demand," commented Jürgen Hatheier, Chief Technology Officer, International at Ciena. "Historically, network traffic has grown at a rate of 20-30% per year. AI is set to accelerate this growth significantly, meaning operators are rethinking their architectures and planning for how they can meet this demand sustainably."

The survey findings underscore the need for enhanced fibre optic capacity, with 87% of participants indicating a requirement for performance levels at or exceeding 800 Gb/s per wavelength to meet anticipated AI demands.

As sustainability becomes increasingly crucial, 98% of data centre experts acknowledged the importance of pluggable optics solutions for reducing power consumption and the physical footprint of infrastructure.

The research also indicated a shift towards more distributed computing frameworks. A significant 81% of respondents believe that training large language models (LLMs) will occur across multiple distributed data centre facilities, necessitating interconnected DCI solutions.

Further preferences were identified regarding AI infrastructure deployment strategies. A majority prioritised AI resource utilisation over time (63%), reducing latency by placing inference compute closer to users at the network edge (56%), and data sovereignty requirements (54%). Strategic location offerings for key customers also featured prominently (54%).

Most respondents, around 67%, prefer Managed Optical Fibre Networks (MOFN) for long-distance connectivity over deploying dark fibre.

"The AI revolution is not just about compute - it's about connectivity," added Hatheier. "Without the right network foundation, AI's full potential can't be realised. Operators must ensure their DCI infrastructure is ready for a future where AI-driven traffic dominates."

The survey, conducted by Censuswide, covered markets including the United States, United Kingdom, Germany, Saudi Arabia, Norway, Sweden, Australia, South Korea, India, Philippines, Indonesia, Brazil, and Mexico. The data highlights a global shift in focus to accommodate AI-driven demand within data centres, underscoring the need for robust and scalable networking solutions.