Google Cloud expands AI infrastructure in India, backs IIT Madras evaluation platform

Google Cloud announced major AI investments in India, including expanded local compute capacity powered by Trillium TPUs and support for a new community-driven platform to evaluate AI models for Indian languages and contexts.

The tech giant is significantly increasing AI hardware capacity in India through its AI Hypercomputer architecture, enabling businesses and public sector organizations to train and deploy advanced Gemini models locally. This expansion addresses data residency and sovereignty requirements while reducing latency for AI applications.


Google Cloud is also providing early access to its most advanced Gemini models to Indian customers with full data residency support—a first for the company. The deployment includes Gemini 2.5 Flash with batch processing capabilities, Document AI for automating document workflows, and grounding features using Google Maps for location-aware responses.

IIT Madras' platform tackles language, cultural gaps in AI benchmarking

In partnership with IIT Madras, Google Cloud is supporting the launch of Indic LLM-Arena, an independent evaluation platform operated by the AI4Bharat center. The platform uses crowd-sourced, blind comparisons where users submit prompts in any Indian language—including code-mixed queries like Hinglish—and vote on responses from anonymous AI models.

"At AI4Bharat, our mission is to build AI for India's specific needs," said Mitesh Khapra, associate professor at IIT Madras. "Indic Arena will be that platform" for standardized benchmarking across India's multilingual landscape.

The initiative addresses critical gaps in current AI evaluation, which predominantly focuses on English and Western contexts. Indic LLM-Arena evaluates models on three dimensions: linguistic diversity including code-switching, cultural appropriateness for India's regional contexts, and safety filters for India-specific harms like communal misinformation and caste-based stereotypes.

The platform currently supports text-based inputs and plans to expand to vision, audio, and agentic tasks involving documents and web searches. Google Cloud is providing cloud credits to power the infrastructure.

The moves underscore India's growing importance in the global AI ecosystem, with Vice President Saurabh Tiwary noting that India's "developer community, vibrant startup ecosystem, and leading enterprises are embracing AI with incredible speed."