E-Paper
Securing our AI future
This article is authored by Sharad Sanghi, CEO & co-founder, Neysa.
India’s digital journey is a story of bold, foundational choices. We built public infrastructure to serve a billion people, moving beyond the adoption of global technologies. Aadhaar created a universal identity layer that unlocked financial inclusion for hundreds of millions. UPI expanded this foundation, creating frictionless rails for value exchange and revolutionising payments at a scale the world had never seen. Now, Open Network for Digital Commerce (ONDC) is set to transform e-commerce, leveling the playing field for small businesses.
This India Stack philosophy of building open, interoperable, and nationally-owned digital backbones has been our greatest strategic asset. It allows private innovation to flourish on our own terms, creating a vibrant ecosystem that is both competitive and resilient.
Today, we stand at the threshold of the next great technological shift: Artificial Intelligence (AI). AI is rapidly evolving into the new intelligence layer of our economy, a cognitive fabric influencing decisions across finance, health care, logistics, and governance. Once again, we face a critical choice: will we rent this intelligence from a handful of global providers, or will we build the capacity to shape our own cognitive future? For a nation with India’s scale and ambition, the answer is clear. We must build.
For the last decade, the default approach for most enterprises has been to consume AI as a service from large international players. This is the black box model: We provide the data and receive an output, but the intricate machinery within, including the model architecture, the training data, and the governance rules, remains opaque and outside our control. What began as convenience has come at a steep and growing price, creating a profound strategic dependency.
This dependency creates several tangible risks. First, it exposes us to algorithmic bias. Models trained mainly on western datasets often fail to capture India’s unique complexities, including its diverse languages, dialects, and cultural contexts. An AI tool that performs well for retail banking in New York might make critical errors when assessing credit risk for a farmer in rural Maharashtra. Second, it leads to immense cost volatility and vendor lock-in. Imagine a fast-growing Indian startup building its entire customer service operation on a proprietary AI API, only to face a sudden 300% price hike that cripples its business model. With no alternatives, there is no leverage.
Finally, and most critically, it presents a geopolitical and supply chain risk. In an increasingly fragmented world, access to critical technology must be secured. Just as nations safeguard energy supplies or semiconductor production, we must now think in terms of intelligence security. Relying on foreign-hosted black-box models for vital infrastructure, from power grids to financial markets, creates a strategic vulnerability we cannot afford. The era of rented intelligence, with all its hidden risks, is no longer sustainable. Fortunately, the rise of powerful open-weight models, fast-growing sovereign models, and increasingly accessible compute infrastructure gives us the tools to forge our own path, to bring clarity, and move beyond the black box.
To secure our digital future, the national mission must evolve from Make-in-India for manufacturing to Build-in-India for intelligence. The first focused on building factories and physical capacity. The second is about creating a sovereign intelligence infrastructure that can power the entire nation's digital economy.
This vision is not about producing a single, monolithic Indian AI. Instead, it is about fostering an open, modular, and vibrant ecosystem defined by choice and flexibility. A sovereign AI control plane would act as a national foundry for intelligence, enabling enterprises to:
Access a curated catalogue of models: Host a range of the world’s best open-weight frontier models alongside powerful models developed in India and trained on local datasets.
Prevent vendor lock-in: Establish interoperability standards that allow businesses to switch seamlessly between model providers, fostering genuine competition on price and performance.
Enable local fine-tuning: Provide tools for an agritech startup to adapt a base model to identify crop diseases specific to Indian climates, or for a hospital to improve accuracy in reading medical scans for the Indian physiognomy.
Ensure data sovereignty: Guarantee that sensitive Indian data is processed and stored within national borders, upholding privacy and security mandates.
This approach transforms AI from a high-cost, high-risk import into a democratised, accessible utility. It lowers barriers to entry and empowers a new generation of Indian startups, researchers, and public institutions to innovate on a secure and sovereign foundation.
The rewards of a Build-in-India AI strategy are immense and multifaceted, creating a virtuous cycle of economic and strategic gains.
Economically, local governance over AI dramatically reduces the total cost of ownership. It frees enterprises from the unpredictable pricing of global monopolies and allows them to build more efficient, context-aware solutions. More importantly, it will spark a high-value domestic AI industry, creating jobs for AI engineers, data scientists, and ethicists, and positioning India as an exporter of specialised AI solutions.
Strategically, the benefits are even more profound. For regulated sectors such as banking and healthcare, a sovereign AI framework provides the ultimate assurance: auditability. Regulators like the RBI can gain full transparency on how a model used for lending decisions was trained and tested, ensuring compliance with Indian laws and ethical guidelines. This builds institutional trust and guarantees that as AI grows in power, it remains accountable.
UPI demonstrated how open digital infrastructure can unlock competition and lower costs. A sovereign AI control plane will extend that model to intelligence itself, fostering a resilient, self-sustaining ecosystem and ensuring India’s digital autonomy for decades to come.
Some worry that a focus on sovereignty could slow innovation. In practice, a well-designed sovereign framework accelerates it. By providing access to pre-vetted models, secure data environments, and easy-to-use fine-tuning tools, it reduces risk and shortens the time required to bring new AI applications to market. Sovereignty, in this sense, becomes a powerful competitive advantage.
Achieving this vision requires a concerted effort across the nation. It is a national mandate:
Policy must champion open standards for interoperability, fund foundational model research, and create sandboxes for responsible innovation.
Industry leaders must look beyond short-term pilots and view sovereign AI as a long-term strategic investment in resilience and competitiveness.
Technology providers must deliver the platforms and infrastructure that make this vision a reality, offering sovereignty as a seamless, high-performance, and accessible service.
Build-in-India for AI is the cornerstone of our nation’s competitiveness, security, and prosperity in the coming cognitive age. The time to build our intelligent future is now.
This article is authored by Sharad Sanghi, CEO & co-founder, Neysa.
