According to a joint survey by Ciena and Censuswide, the rapid growth of artificial intelligence (AI) workloads is driving a major transformation in data centre network infrastructure, with global data centre experts anticipating a significant increase in interconnect bandwidth needs over the next five years.
The survey queried more than 1,300 data centre decision makers across 13 countries. More than half (53 per cent) of respondents believe AI workloads will place the biggest demand on data centre interconnect (DCI) infrastructure over the next two to three years, surpassing cloud computing (51 per cent) and big data analytics (44 per cent).
To meet surging AI demands, 43 per cent of new data centre facilities are expected to be dedicated to AI workloads. With AI model training and inference requiring unprecedented data movement, data centre experts predict a massive leap in bandwidth needs. In addition, 87 per cent of participants believe they will need 800 Gb per second or higher per wavelength in terms of needed performance of fibre optic capacity for DCI.
Respondents also mentioned that there is a growing opportunity for pluggable optics to support bandwidth demands and address power and space challenges. According to the survey, 98 per cent of data centre experts believe pluggable optics are important for reducing power consumption and the physical footprint of their network infrastructure.
The survey found that, as requirements for AI compute continue to increase, the training of Large Language Models (LLMs) will become more distributed across different AI data centres. It highlighted that 81 per cent of respondents believe LLM training will take place over some level of distributed data centre facilities, which will require DCI solutions to be connected to each other.
Furthermore, the key factors where AI inference will be deployed rank the following priorities:
- AI resource utilisation over time is the top priority (63 per cent).
- Reducing latency by placing inference compute closer to users at the edge (56 per cent).
- Data sovereignty requirements (54 per cent).
- Offering strategic locations for key customers (54 per cent).
Moreover, rather than deploying dark fibre, the majority (67 per cent) of respondents expect to use Managed Optical Fibre Networks (MOFN), which utilise carrier-operated high-capacity networks for long-haul data centre connectivity.
Commenting on the survey, Jürgen Hatheier, chief technology officer, International, Ciena, said, “AI workloads are reshaping the entire data centre landscape, from infrastructure builds to bandwidth demand. Historically, network traffic has grown at a rate of 20-30 per cent 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.”
He further mentioned, “The AI revolution is not just about compute, it is about connectivity. Without the right network foundation, AI’s full potential cannot be realised. Operators must ensure their DCI infrastructure is ready for a future where AI-driven traffic dominates.”
