Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Economic Survey proposes ‘AI-OS’ platform to make AI a public good in Simple Termsand what it means for users..
“This will enable the sovereign to collaborate with state and local institutions to expand the availability of structured, anonymised, and machine-readable datasets in priority sectors,” the Survey said.
The proposed initiative aims to address fragmentation in data availability, quality, standardisation and interoperability, which has so far limited the scalability of domestic AI innovation.
“For such a decentralised innovation ecosystem to translate into tangible outcomes, deliberate coordination is essential,” the report added. While industry experts said the proposal could significantly expand access to AI tools, the implementation should be in an open and inclusive manner.
Jaspreet Bindra, founder of AI&Beyond, said the announcement was expected ahead of the upcoming AI summit and could help democratise access to AI if executed well.
“If it becomes too closed or controlled, it won’t work. It has to follow the same model as DPI and DPG, where everyone can participate and build on top of it. That’s what made UPI and Aadhaar successful,” he said.
Another startup founder, Raveen Sastry, who runs the AI lab and launch studio Basethesis, said access to both data and computing power will determine the success of the platform. Along with data, computing is one of the two most critical components in the whole proposal, he said.
“It’s great to talk about shared compute, but someone has to pay for it. So that’s a practical challenge. Still, if it is done right, access to shared compute can be hugely beneficial for startups and researchers,” Sastry added.
Focus on application-led AI
The Survey said India should adopt a bottom-up approach to AI development, focusing on application-specific solutions rather than competing directly in building large, capital-intensive foundational models.
Globally, AI development has split into two paths, with Western firms focusing on frontier models backed by massive capital and computing power, while other regions follow decentralised, application-driven models, it said.
Given limited access to advanced computing infrastructure and funding, India is better placed to focus on smaller, targeted AI models for sectors such as healthcare, agriculture, finance, education, and governance. “These models are more efficient, easier to fine-tune, and better suited to India’s existing infrastructure,” the report noted.
The Survey highlighted that India’s scale and digital diversity give it a strong advantage in developing AI-based applications and services. This makes India’s database strategically important in the global AI race.
It suggested that companies generating significant revenues from Indian datasets may be expected to contribute to the domestic AI ecosystem. “Value extraction must be accompanied by value creation in India,” the survey said.
Srikanth Velamakanni, cofounder and group CEO of Fractal, said India should use the proposed AI-OS platform to prioritise sector-specific deployment instead of building a one-size-fits-all system. “If we deploy AI at scale in sectors like healthcare, we will start generating massive amounts of high-quality data. There will be inferences, feedback loops, and continuous learning happening in real-world settings,” he said.
“Even if today we don’t have the best model — even if we can’t beat something like GPT-5 — within a couple of years, we would have collected so much real-world data that our models could become significantly better. That is India’s advantage — scale,” Velamakanni said, adding that building domestic models backed by strong feedback loops would be key to winning the AI race.