Transforming network strategy to meet new AI demands

The world’s communications industry is at a crossroads. Demand for bandwidth is reaching historic highs, with 800G becoming mainstream to support the growth of AI and the move to 1.2T and 1.6T looming into sight.

But for organisations across the globe, increasing capacity is now much more than just linear scaling to meet basic connectivity and application needs.

With AI and cloud growth driving a massive rise in demand for improved interconnectivity and latency, while stricter data sovereignty requirements and consolidation change the shape of the market, a rethink is called for in the provision of network capacity at a scale that legacy architecture was not built to support.

“Workloads have become unpredictable and large, and the conversation is no longer adding incremental capacity,” says Kapil Kumar Jain, vice president for global network services at Tata Communications. “It’s building intelligent, scalable and resilient digital infrastructure that can support the decade of innovation.”

He illustrates the challenge by referring to the comfortably double-digit annual growth that Tata is seeing across major international routes, with cloud adoption, content distribution and data centre expansion acting as key drivers alongside the rise of AI.

More distributed data centres, and the separation of AI training and inference workloads, are creating a requirement for further changes in the network landscape.

Despite this need for transformation, organisations are often lagging behind trends in the rapidly changing environment. Nearly two-thirds of global enterprises surveyed by Bloomberg in a recent Tata-commissioned study continue to operate on legacy infrastructure, risking missing out on huge amounts of revenue from the AI opportunity.

However, organisations can alter this dynamic by partnering with a connectivity partner that has a sufficient network scale, capacity and forward-thinking stance to capitalise on the rapid growth in the market.

Those that look set to succeed are global carriers that take a multi-layered approach, including deploying higher-capacity optical technologies, upgrading terrestrial and subsea routes, expanding data-centre-interconnection ecosystems and enhancing network automation.

Jain adds that one pathway to seizing on the opportunity is collaborating more closely with international partners to improve the economics of infrastructure and ramp up the speed of construction.

He cites moves such as Tata’s recently announced investment as a consortium member in a new subsea cable system connecting Chennai in south-eastern India with Singapore – expected to be ready for service in late 2029.

At the same time, the company is investing to integrate and expand its presence on a subsea cable connecting India, Malaysia, Singapore, Myanmar and Thailand, having last year integrated the TGN-IA2 system constructed by the Asia Direct Cable consortium into its network fabric.

Together, these investments are expected to help fulfil the AI-driven requirements of enterprises both across Asia and globally, with Tata Communications seeing strong demand on routes connecting India with Southeast Asia, Europe, the Middle East and North America.

“India is emerging as an important hub due to increasing investments in cloud infrastructure, AI innovation and digital services,” says Jain.

Automating connections

In addition to the company also investing in 800G, it is enhancing automation through moves like the launch earlier this year of a software-defined platform known as IZO Data Centre Dynamic Connectivity.

This is aimed at transforming how enterprises connect their data centres so they can better manage fluctuating AI workloads, operating as a self-healing intelligent network covering key global data centres across five continents.

The platform can autonomously reroute traffic within seconds during disruptions, enhancing resilience by protecting against costly cable outages and downtime, while helping deliver predictable latency and performance. At the same time, enterprises can use the service to monitor connectivity performance, receive proactive alerts and dynamically scale bandwidth to meet changing needs.

“Future network upgrades will increasingly be enabled through software-defined control, optical innovations and automation,” says Jain. “By continuously investing in network modernisation and route diversity, we can reduce the time required to deliver additional capacity to customers.”

For organisations that need broad and reliable support for international operations, a global provider like Tata is ideally place to serve their needs – with the company having more than 650,000km of subsea and 140,000km of terrestrial fibre.

“Scale matters more than ever in the AI era,” says Jain. “A provider with extensive global reach, diversified subsea and terrestrial assets, and direct connectivity into major cloud, internet and data centre ecosystems is far better positioned to absorb traffic growth and support future demand.”

In addition, such scale reduces the operational complexity of dealing with multiple providers, while enabling the company to look ahead with authority when assessing the global requirements created by new AI-driven realities.

“Our long-term investment philosophy enables us to build ahead of demand,” says Jain. “This ensures that customers are prepared for future capacity requirements rather than simply addressing current needs.”

Tata Communications integrates new TGN-IA2 subsea cable into global network