At this year’s OFC conference, discussions around optical interconnects in AI scenarios—particularly for scale-out and scale-up architectures—were exceptionally active. Dozens of related workshops and panel discussions took place (as shown in the image below). Interestingly, some experts presented the same slides in different sessions, repeating their points despite differing stances, which made for an engaging dynamic. Here’s a curated summary of key insights for your reference.

It has been a while since my last blog, as I was busy preparing for and attending OFC. Afterward, I needed time to digest all the valuable insights I gathered. To be honest, getting started on this blog was daunting—I wasn’t sure which topic to address first, given the wealth of fascinating discussions at OFC. Unlike previous years, the focus wasn’t primarily on broadband; instead, AI connectivity took center stage, along with its vast bandwidth and low latency requirements.

The rapid advancement of Artificial Intelligence (AI) and Machine Learning (ML) is fundamentally reshaping datacenter network architectures. Presentations at the OFC Summit on Optics for AI Datacenters highlighted a clear trajectory driven by the insatiable demand for data. Three interconnected themes emerged: the sheer scale of future connectivity needs, the ongoing shift from traditional pluggable optics to Co-Packaged Optics (CPO), and the longer-term evolution towards highly integrated Optical I/O (OIO).

Distributed sensing at the access level of the FTTx network could have several benefits, such as notification when digging is done to close to the fiber, when there is tampering with the fiber, structural health of the cable or sensing of geological events (flooding, earth slide, etc…).
However, the cost of full-on distributed sensing is not cost effective, but there are a few alternatives, these techniques leverage standard or specialized optical equipment and analysis of light propagating through the fiber to detect physical changes. They generally offer lower initial cost than dedicated DS systems but may have limitations in spatial resolution or the type of information they provide.  

Rodrigo Amezcua Correa, Relativity Networks, USA
Paolo Dainese, Corning, USA
Russell Ellis, Microsoft, United Kingdom
Kerrianne Harringtone, University of Bath, United Kingdom
Matěj Komanec, CTU, Prague, Czech Republic
Andrew Lord, BT, United Kingdom
Kazunori Mukasa, Furukawa, Japan
Mohammad Pasandi, Ciena, Canada
Pierluigi Poggiolini, Politecnico di Torino, Italy
Yingying Wang, Linfiber and Jinan University, China
YOFC
China Telecom
Sumitomo (Sato)

This report summarizes key insights and trends from several OFC conference sessions, highlighting major shifts in network architecture, optical technologies, and strategic opportunities driven primarily by the demands of AI and escalating bandwidth requirements.

How are you seeing FTTx adapting to meet the demands of automation and digital equality? My latest blog examines how configuration flexibility and adaptability, especially the shift to direct distribution, is playing a crucial role

Is Your RAN Cable Infrastructure Ready to Unlock the Monetization Potential of AI? The other day I watched the webcast of Light Reading on The Role of AI in Platforms for Future RAN Systems [1], with NVIDEA, Fujitsu and Supermicro sharing their thoughts and market vision. Below you will find my interpretation of this webcast. The telecommunications industry is rapidly evolving, and the integration of Artificial Intelligence (AI) in Radio Access Networks (RAN) is at the forefront of this transformation.

Maximizing Space: How Reduced Cladding and Coating Fibers Improve Data Center Efficiency Just the other day, I was reminiscing about some exploits from the past. One that came to mind was when I had to put together a mock-up to analyze how much Cat6a we could responsibly manage in an 800mm wide 42RU rack. At that time, we established that anything beyond 240 patches would significantly increase the operational risk during moves, adds, and changes. It was also around that

In the first three parts of this blog series, we explored various critical aspects of IT cabling for AI clusters, emphasizing the impact of high bandwidth and latency requirements on connectivity solutions, the practical considerations of deployment and installation, and the sustainability implications of different cabling choices.