Asia’s push to expand artificial intelligence infrastructure is increasing the need for resilient undersea cables, with damage to the networks carrying international data capable of disrupting cloud services, financial transactions and supply chains.
The warning comes as businesses invest heavily in computing capacity and data centres, while the physical networks connecting them face growing and increasingly unpredictable demands. More than 99% of international data is carried by subsea cables rather than satellites.
Recent incidents have highlighted the risks. Cable damage in 2024 caused widespread internet outages across Africa, affecting banks, fintech platforms and cloud-based businesses. Cuts to cables in the Red Sea in 2025 severely disrupted access across Asia and the Middle East.
Traffic was forced on to longer routes, increasing latency for cloud services and adding costs for companies whose operations depend on the rapid movement of data.
AI workloads and hyperscale data centres are also driving demand for greater bandwidth. Connections operating at 400 gigabits per second are described as a baseline, while 800 gigabits per second is already emerging.
That is putting pressure on companies and governments to consider the networks supporting AI alongside investment in computing power. Without adequate subsea cables and end-to-end connectivity, expensive AI systems could face disruption, higher operating costs and unreliable access to data.
Singapore and India emerge as key data transit hubs
Singapore and India are becoming important transit points for regional and international data. Singapore acts as a gateway to south-east Asia and has one of the world’s densest concentrations of data centres.
India links Asia with the Middle East and Europe through major cable landing points and expanding data-centre capacity in cities including Mumbai and Chennai. Their strategic positions make diverse and resilient routes between the two countries and other financial and commercial centres increasingly important.
The network architecture supporting those hubs will need to evolve, moving beyond infrastructure designed for predictable, linear traffic. AI-driven demand is more volatile, requiring flexible connections between data centres that can expand rapidly as usage changes.
That means coordinating subsea cables with terrestrial fibre and individual data centres. A bottleneck at a coastal landing station can undermine the capacity of a fast transoceanic cable, while poor integration between systems can limit the benefits of investment elsewhere in the network.
Resilience also depends on capacity planning, route diversity, protection and repair readiness, as well as long-term operational management. Monitoring systems, strategically held spare equipment and clear governance can help prevent an individual cable fault from developing into a wider regional outage.
Public policy is beginning to reinforce those measures. The Enhanced ASEAN Guidelines for Strengthening Resilience and Repair of Submarine Cables recommend limiting repair processing times to 10 days, making faster restoration part of efforts to protect economic stability.
A Tata Communications study found that AI investment is a strategic priority for 77% of enterprise leaders, while 65% of enterprises continue to rely on legacy or still-developing infrastructure.
For businesses, the issue is whether connectivity receives the same attention as computing capacity and data centres. Cables, landing stations and alternative routes may be less visible in investment decisions, but underinvestment in them could leave AI infrastructure exposed to congestion, outages and rising costs.
As AI adoption accelerates, the resilience of the networks beneath the sea will remain central to Singapore’s ambitions as a global digital hub and to the wider stability of Asia’s digital economy.
