Google's $15B Data Centre Expansion in India: Scaling AI Infrastructure in Asia

Why Google is building more data centres in Asia

Google is initiating construction on a US$15 billion data centre hub in Andhra Pradesh starting April 28, 2026. This 1 gigawatt project, spanning three campuses and over 600 acres, signals a strategic pivot toward Asian infrastructure. The expansion aims to address both surging AI compute requirements and regional data residency mandates.

Why This Matters

The technical reality of artificial intelligence necessitates moving compute capacity closer to the edge. While model training can occur in centralized hubs, real-time AI inference and low-latency digital services require infrastructure to be physically located near the end-user. As digital activity surges in regions like India and Indonesia, the legacy model of relying on distant US or European hubs becomes unsustainable due to latency constraints and performance bottlenecks.

Furthermore, the shift is driven by a hardening regulatory landscape. Governments are increasingly enforcing data localization frameworks that require sensitive data to remain within national borders. For hyperscalers like Google, building local 1GW-scale facilities is no longer just an optimization—it is a prerequisite for operating in high-growth markets while maintaining compliance with regional data protection laws.

Key Insights

• Google is developing a 1GW data centre hub across three campuses in Adavivaram, Tarluvada, and Rambilli (The Economic Times, 2026).
• Global data centre electricity use is projected by the International Energy Agency (IEA) to more than double by 2030, driven specifically by AI power density needs.
• Data residency regulations in India, Indonesia, and Vietnam are forcing cloud providers to localize storage and processing infrastructure.
• AI workloads are shifting data centre design toward higher power density and specialized cooling for GPUs and hardware accelerators.
• The Andhra Pradesh cluster will utilize subsea cable links to connect directly with global network nodes in Singapore, Malaysia, and Australia.
• Google currently operates data centres across 29 global sites, with the new Visakhapatnam cluster set to become its largest in Asia.

Practical Applications

• Use Case: Indian enterprises deploying real-time AI inference models on the 1GW Andhra Pradesh hub to minimize latency for local users. Pitfall: Overestimating the immediate availability of power at scale, which can lead to phased deployment delays.
• Use Case: Multinational corporations utilizing Google’s localized infrastructure to comply with India’s evolving data protection and localization frameworks. Pitfall: Implementing rigid data residency without a hybrid cloud strategy, resulting in fragmented data silos.
• Use Case: Hyperscale expansion through local partnerships with entities like AdaniConnex and Bharti Airtel to secure land and power grid access. Pitfall: Complexity in multi-party implementation leading to inconsistent infrastructure standards across different regional campuses.

References:

• https://www.cloudcomputing-news.net/news/why-google-is-building-more-data-centres-in-asia/