As artificial intelligence (AI) and cloud computing continue their relentless march forward, the digital infrastructure underpinning these technologies is experiencing unprecedented demands. At the recent Pacific Telecommunications Council conference (PTC’25) in Honolulu, Hawaii, industry leaders convened to discuss the critical role of fibre in meeting these challenges and shaping the future of global connectivity.

The session, which brought together executives from major players in the telecommunications and cloud computing sectors, painted a picture of an industry on the cusp of a transformative era. Steve Smith, CEO of Zayo, set the tone by drawing parallels with the early days of the cloud revolution. “Everything that is going on is similar to 20 years ago, the beginning of the cloud cycle,” Smith remarked. “This [AI] will be put everywhere, manufacturing floors, airports, hospitals.”

Power: The New Currency of Digital Infrastructure

A recurring theme throughout the discussion – and the conference indeed – was the paramount importance of power in the evolving digital landscape. Lynn Smullen, Senior Vice President of North America Cloud Infrastructure at Oracle, said: “Digital infrastructure is the heart of it all, and power is the source.”

The increasing power requirements of AI and cloud computing are reshaping deployment strategies. Smith noted that “deployments will go where the power sources are,” noting a shift in focus from traditional metro areas to locations with abundant energy resources. This trend is driving significant infrastructure investments, with Zayo alone adding approximately 12,000 miles of fibre routes last year and allocating close to US$100 million in capital to support these expansions.

Dave Schaeffer, Founder and CEO of Cogent Communications (NASDAQ: CCOI), underscored the unprecedented scale of investment in the sector. “More than US$2 trillion announced for deployment in the last couple of years will not bankrupt those companies,” he stated, distinguishing the current boom from the dot-com bubble of the early 2000s.

The AI Factor

The advent of AI is not just increasing demand; it is fundamentally altering the nature of data processing and storage. Schaeffer pointed out that “current data sets are not comparable, and the power is harder to find,” highlighting the unique challenges posed by AI workloads. The massive data requirements for AI training are driving “tremendous demand for transporting networks,” as the zettabytes of stored data become the raw material for AI systems.

Gary Smith, President and CEO of Ciena (NYSE: CIEN), observed a significant uptick in cloud traffic, particularly among hyperscalers. “We are seeing the first investment signs of that cascading through,” Smith said, referring to the US$2 trillion investment in GPU infrastructure and its ripple effects on power and network infrastructure.

Financing the Future

The scale of investment required to meet these infrastructure demands is prompting a re-evaluation of financing models. Smullen touched on the challenge of forecasting power requirements and called for more creative funding solutions.

Radhika Venkatraman, Senior Managing Director at Cerberus Capital Management, offered insight into the investment landscape. “We are looking at an ecosystem perspective,” she explained, detailing how her firm is examining opportunities across the entire value chain, from cell towers and fibre to data centres and the components within them.

The Next Decade

Steve Smith of Zayo provided a bullish outlook for the industry’s future. “Over the last 10 years, some US$5 trillion was spent on realising the public cloud. I think it will be US$5-6 trillion to make AI real in the next decade,” he predicted. This massive investment is expected to drive innovation and growth across the entire digital infrastructure ecosystem.

Despite the optimistic projections, the panel acknowledged several challenges facing the industry. Data sovereignty concerns are influencing traffic patterns, with much AI-related data remaining within national borders. The integration of siloed systems, particularly in the telecommunications sector, was identified as a hurdle for AI implementation.

Venkatraman raised questions about the proliferation of large language models (LLMs), suggesting that consolidation may occur as enterprises begin to explore AI applications more deeply. “The question now is if we need a million different models,” she said.

An Invisible Revolution

As the industry gears up for this next phase of digital transformation, taking to the stage at the conference, Sylvia Luke, Lieutenant Governor of the State of Hawaii, offered a poignant reminder of the sector’s unique position. “The less visible you are, the better you work, and because of that, you will never get the praise you truly deserve,” she said, acknowledging the critical yet often unrecognised role of digital infrastructure.

The consensus among the panellists was clear: the world is on the brink of what could be categorised as the fourth industrial revolution. As Steve Smith of Zayo urged, “I think every person in this room needs to study it, read more about it, prepare for it.”