India to Unveil Voice-Based Large Language Model Ahead of Delhi AI Summit

Washington – India plans to launch a new voice-enabled Large Language Model (LLM) ahead of next year’s global Artificial Intelligence (AI) summit in New Delhi, according to Abhishek Singh, a senior government official, who outlined an ambitious roadmap to accelerate the country’s AI capabilities.

Singh, Director General of the National Informatics Centre and Additional Secretary at the Ministry of Electronics and IT, said India’s rapid growth as the world’s fastest-growing major economy is closely linked to its robust digital infrastructure.

“We are recognized as the world’s largest, fastest-growing economy… growing at 8.2 per cent,” he said, describing digital public infrastructure—including Aadhaar, UPI, DigiLocker, and India Stack—as “the basic foundation” for India’s AI ambitions.

India aims to scale its economy from around $4 trillion to $30 trillion by 2047, with AI playing a key role in enabling services in all Indian languages, improving access to government schemes, healthcare, and agriculture support for millions.

Singh highlighted previous gaps in India’s AI ecosystem, noting the country had only around 600 GPUs and limited research spending compared to the U.S. and China. The government has since facilitated access to 40,000 GPUs at subsidized rates and is funding nearly a dozen initiatives to develop indigenous LLMs and SLMs, including sector-specific models for healthcare and materials science.

Two projects led by IIT Madras and IIT Bombay are nearing completion. “Before the summit… we should be able to announce an Indian LLM, primarily a voice-based LLM,” Singh said.

India is also building a national datasets platform—AI Coach—currently hosting 3,500 public and private datasets to support AI innovation. The country is developing 30 scalable AI applications, including an AI assistant for farmers and diagnostic tools for tuberculosis, cataracts, and diabetic retinopathy.

Talent development is a key pillar of India’s AI strategy, with funding for research, data labs in ITIs and polytechnics, and support for students at undergraduate, postgraduate, and doctoral levels. AI safety remains central, with the establishment of an AI Safety Institute and tools for bias mitigation, privacy preservation, deepfake detection, and labeling AI-generated content.

The AI Impact Summit 2026 will be the first global AI gathering hosted in a developing country, with over 100 countries, 15 heads of government, and 50 CEOs expected to attend. Outcomes will include an AI charter for democratized access, an AI commons repository, workplace transition principles, and commitments from frontier model developers to share usage data with sovereign governments.

Global tech leaders from companies including Anthropic, Zoom, Microsoft, Salesforce, OpenAI, Equinix, and NVIDIA praised India’s growing centrality in AI, while UN Special Envoy Amandeep Singh Gill emphasized the need for inclusive multilateral engagement. Google executive Ben Gomes highlighted India’s scale and potential, emphasizing AI’s role in enhancing, not replacing, human learning.

With inputs from IANS