Cross-Border Fiber Market Forecast to Reach $36.82bn by 2031 as Distributed AI Drives Demand

The global cross-border fiber-route market is projected to expand at a 10.21% CAGR between 2026 and 2031, reaching $36.82bn by 2031, according to estimates attributed to Mordor Intelligence in a September 2026 industry review. The same review places the market at $22.65bn in 2026, following a $20.13bn valuation in 2025.

The stated demand driver is the growth of AI clusters and GPU deployments distributed across countries and regions. These architectures require high-capacity, low-latency interconnection between remote data-center locations, shifting transport-network availability closer to the core infrastructure considerations that already include power, land and grid access.

The review also points to greater direct hyperscaler involvement in physical cable assets. Hyperscalers accounted for 75% of international subsea-cable capacity in 2025, according to the cited data, and were directly involved in more than two-thirds of major planned cable-laying projects. That participation is presented as part of a broader move toward direct control of fiber resources rather than reliance solely on conventional capacity procurement.

Delivery timelines are lengthening. The article says the period from procurement to completed cable installation had extended from 18 months to 48 months by 2026. Standard cable-landing procedures may take six to 12 months, while approvals for routes through environmentally protected areas may require 24 to 36 months. Separately, individual Red Sea incidents required up to eight weeks to deploy repair vessels; this is an example rather than a general repair benchmark.

The review identifies several procurement and route-design models expected to grow through 2031. The market for dark fiber and long-term fiber-pair rights of use, or IRUs, is forecast to grow at an 11.86% CAGR between 2026 and 2031. Neutral wholesale procurement is projected to grow at 11.37% annually, while hybrid terrestrial-subsea backbones are forecast to expand at an 11.59% CAGR.

For operators supporting geographically distributed AI workloads, these trends could make controlled, redundant and scalable interregional transport networks a stronger competitive differentiator than local site capacity alone. Given the longer delivery cycles, early acquisition of IRUs, dark fiber and route diversity could become an important element of AI infrastructure siting and deployment strategy.