AI democratization crucial for India’s global leadership, say experts at Mint AI Summit 2025 | Mint

India must democratize AI to establish itself as a global leader in artificial intelligence, AI experts deliberated at the Mint AI Summit 2025. They highlighted the need for collaboration across multiple sectors, including public-private partnerships, universities, and scholars.

Currently, India lags in AI PhD holders and faces significant infrastructure challenges, particularly in energy supply and scaling AI development. Experts stressed that building a robust AI ecosystem requires thoughtful planning, workforce development, and infrastructure expansion.

“The democratization of AI is essential for it to be taken seriously,” said Prasad Balakrishnan, COO of MiPhi. “Universities, scholars, the government, and enterprises all have vital roles to play in building a strong AI ecosystem.”

Metesh Bhati, chief digital and AI officer at Protean eGov Technologies Ltd, urged the government to open datasets to startups, enabling them to train AI models and unlock AI’s full potential. “If the India Council for Agricultural Research opens up anonymized data for startups, it could fundamentally transform the landscape. India is an agri-first country, and such advancements could be huge,” Bhati said.

Also read | Well-equipped data protection board a must to enforce digital privacy laws, say experts

Experts also underscored the importance of skill development to foster AI innovation. “One of the biggest challenges in India’s AI mission is skill development. We need to invest in building a skilled workforce to drive this technology forward,” Balakrishnan added.

Manu Jain, chief executive officer (CEO) of G42 India, emphasized the need for collaboration among universities, research institutes, government, and private entities. “If we build strong collaborations and fund the necessary compute infrastructure, we can make this happen. AI development requires cooperation between the public and private sectors, and this will be key to India’s success.”

The Economic Survey 2024-25 highlighted that India’s workforce in low-skill and low-value-added services remains vulnerable to AI disruptions. It recommended creating “robust institutions” to help workers transition to medium- and high-skilled jobs, where AI can augment rather than replace them.

The experts also advised focusing on developing foundational and narrow AI models for homegrown innovation. “We might want to work on our own foundation model for our specific needs. Many existing models are not fine-tuned for India. While this should be a long-term goal, we must carefully plan and strategize,” said Vishal Kanvaty, CTO of NPCI.

Also read | Hiranandani’s AI infra firm Yotta plans a US listing. But at what cost?

Kanvaty also raised concerns about dependency on a single GPU supplier, emphasizing the need for alternative, homegrown solutions. “To ensure long-term AI stability, India must develop its own compute infrastructure to avoid potential setbacks,” he said.

Manu Jain suggested prioritizing existing models before developing foundational ones. “The first step is building applications specific to India and training current models for local languages like Hindi, English, and Tamil. Western models often fail to grasp colloquial meanings,” he said. He also recommended establishing a separate AI ministry, similar to the UAE’s approach.

Meanwhile, the Union government is set to launch the IndiaAI Compute Portal to enable stakeholders, including central ministries and state governments, to request compute capacity. The IndiaAI Compute Pillar has sent a memo to all Union ministries, departments, and chief secretaries about subsidized pricing for compute capacity, network, and storage services. Reportedly, the IndiaAI Mission will cover around 40% of computing costs for eligible users.

To accelerate AI-driven innovation in India, Yotta Data Services has been empanelled under the IndiaAI Mission to enhance the country’s AI capabilities. The company will provide over 50% of the AI Mission’s GPU computing capacity, which will be accessible via the upcoming IndiaAI Mission portal. 

Also read | Why big tech is fussing over open-source AI models – it’s all about ownership, privacy, security

Yotta has pledged more than 9,216 GPUs, including 8,192 Nvidia H100 GPUs and 1,024 L40S GPUs, which will be deployed in phases from its NM1 data center. Yotta’s offerings will include AI Labs for students, AI Workspaces, GPU-as-a-Service, and API endpoints for AI models.

The Indian government launched the IndiaAI Mission on 7 March 2024 to bolster the country’s global leadership in AI and ensure AI benefits reach all segments of society. To achieve this vision, the IndiaAI Mission has introduced seven key pillars to strengthen the domestic AI ecosystem.

According to a report by the World Economic Forum, AI has the potential to contribute $500 billion to India’s economy by 2025, revolutionizing key sectors such as agriculture, healthcare, urban planning, and manufacturing.

Despite India intensifying its focus on artificial intelligence (AI), the country represents only 3% of early-stage AI infrastructure and foundational startups. However, AI application-focused startups are growing rapidly, capturing 65% of the market share, according to the ‘Sense AI Annual Ventures 2025’ report by Sense AI Ventures. AI tooling startups make up 22% of the total. 

The report, based on a survey of 849 startups across India conducted last year, highlights that Bengaluru has become the leading AI hub in India, attracting over 40% of the $1.35 billion in funding AI startups received in 2024. Delhi NCR and Mumbai ranked second and third, respectively.

And read | What explains TCS, Wipro, Infosys, TechM flocking to Nvidia