Inside Google Cloud's US$15bn AI Connectivity Investment

Telecommunications infrastructure is emerging as a fundamental enabler for AI deployment at scale, with Google Cloud's US$15bn (£11.8bn) investment programme demonstrating how global connectivity could determine which regions participate in the AI revolution and which risk exclusion.

The America-India Connect initiative represents a five-year commitment to expand subsea cable systems and fibre networks across Asia, Africa, Australia and North America. By positioning India as a central interconnection hub, the programme illustrates how telecommunications infrastructure is becoming inseparable from AI democratisation efforts.

Google collaborates with the Indian government's Karmayogi Bharat programme, which provides cloud infrastructure for the Integrated Government Online Training (iGOT) platform. The system delivers online training to more than 20 million public sector workers across more than 800 districts in 18 Indian languages, demonstrating the scale at which AI services depend on robust telecommunications backbone.

Subsea networks create redundancy for AI workloads

Google has confirmed a new subsea gateway in Visakhapatnam, alongside three subsea routes linking India to Singapore, South Africa and Australia. These routes will directly support connectivity between Asia, Africa and Australia, while linking onward to North America through existing systems.

Four terrestrial fibre-optic routes will complement these subsea investments by connecting the US, India and regions across the Southern Hemisphere. The infrastructure design prioritises redundancy, a critical requirement for AI workloads that cannot tolerate service interruptions.

On India's east coast, a direct fibre path will link Vizag to Chennai and extend towards South Africa. When combined with the Equiano and Nuvem subsea systems, this creates a redundant high-capacity route, with "redundant" meaning the route will be duplicated for reliability. This could ensure that AI training and inference traffic will continue to flow if one path fails.

Brian Quigley, VP of Global Network Infrastructure at Google Cloud, says: "We're as committed as ever to partnering with communities, businesses and governments to foster AI innovation for all, and we are excited to support the next generation of global connection."

A second east coast route will connect Vizag to Singapore, which will integrate with the Bosun and Tabua systems and form a South Pacific path linking North America to India through Australia. These interlinked systems create route diversity, which reduces reliance on single cables and lowers the risk of disruption to AI services that require consistent data transfer speeds.

Vizag is rapidly emerging as a major subsea landing point, adding to Mumbai and Chennai. The additional landing stations could improve traffic distribution and increase resilience across networks carrying AI model parameters and training datasets.

Mesh architecture supports distributed AI infrastructure

On India's west coast, Google plans to construct a direct fibre path between Mumbai and Western Australia. This will integrate with the TalayLink and Honomoana subsea systems, forming another South Pacific route that connects North America and Australia with India.

This new route adds to the Blue, Raman and Sol subsea cables, that together establish a corridor from North America through the Red Sea to Mumbai. Such corridors act as high-capacity data pathways that support cloud computing and AI model deployment across geographically distributed infrastructure.

The network design follows a mesh structure rather than a single linear route. With multiple paths, Google builds infrastructure that could support the demands of large language models and machine learning systems that require massive bandwidth for both training and inference stages.

The approach addresses increasing demand for bandwidth as AI applications continue to scale.

Supporting AI workloads

According to the infrastructure design, reliable telecommunications networks support cloud platforms and data processing capabilities that underpin AI deployment.

High-capacity subsea cables and fibre routes allow faster and more stable connections, which are essential for AI workloads that depend on real-time data access. Model training, distributed computing and edge AI applications all require low-latency connections that traditional infrastructure might not support at scale.

The digital infrastructure stands to enhance India's iGOT platform by improving search and accessibility for millions of workers through AI tools.

Overall, the initiative reflects how telecommunications capacity has become a prerequisite for regions seeking to deploy AI services, with international route resilience increasingly determining which markets can support the computational demands of advanced AI systems.