The company that wires AI data centres is raising prices because hyperscalers have no alternative
The business move
Fujikura, the Tokyo-based fibre-optic cable manufacturer, is raising prices on the specialized cables that link servers inside AI data centres. CEO Naoki Okada told Bloomberg the company expects to surpass its own revenue forecast due to strong, ongoing demand from almost every major US hyperscaler. With no alternative suppliers able to match Fujikura’s quality and capacity, the company is leveraging its strategic position to increase prices.
Why it matters
Fujikura’s cables are a critical component in AI data infrastructure, connecting the thousands of servers that power large-scale machine learning workloads. Hyperscalers rely heavily on these high-quality fibre-optic cables to sustain the ultra-fast data transfer speeds AI workloads require. Because there are effectively no substitutes that meet the stringent technical standards, Fujikura has pricing leverage that operators must absorb. This means infrastructure costs for AI data centres are rising at a time when demand for capacity and performance is surging. The price hike adds another layer of expense for builders and cloud providers already coping with expensive GPUs, chips, and electricity bills.
Who gains and who gets squeezed
Fujikura is the clear winner, converting sustained demand and supply constraints into higher revenues and better margins. Hyperscale cloud providers and AI service companies face increased infrastructure costs with fewer alternatives to mitigate price rises. This could put subtle upward pressure on prices for AI compute services or slow down some expansion projects that depend on tight hardware budgets. Smaller data centre operators and AI startups without direct access to favoured suppliers like Fujikura may find their cost structure squeezed, making it harder to compete.
What to watch next
Track whether Fujikura’s price increases trigger competitors to invest and close the quality or capacity gap in fibre-optic cables. Watch hyperscalers’ infrastructure spending announcements for signs of changed procurement strategies or delayed expansions. Also monitor if rising cable costs lead hyperscalers to explore alternative designs or network architectures that reduce reliance on expensive internal connections. Finally, keep an eye on whether higher data centre infrastructure costs begin to flow through to AI service pricing, potentially impacting the broader AI ecosystem’s growth trajectory.
AI Quick Briefs Editorial Desk
