Home Artificial intelligence AIIndia Should Pace The Deployment, Not The Frontier – OpEd

India Should Pace The Deployment, Not The Frontier – OpEd

by BollywoodNewsAndMovie


Key Takeaways:

  • India signed the San Francisco Declaration under Pax Silica on 15 November 2026, yet the author argues that a developing economy should not simply join the U.S.–China race to frontier models but instead decide sector by sector how fast and how deeply to integrate AI.
  • Practical priorities include rapid, low-stakes adoption in public services and language tools, accelerated deployment in agriculture, healthcare, and disaster response, and more cautious use in defense and cybersecurity where failures carry irreversible costs.
  • With private AI investment still far below U.S. and Chinese levels, uneven data-center capacity, and local environmental concerns, India’s path is framed as “AI for All”—matching speed to domestic needs, infrastructure, and social acceptance rather than chasing exclusive frontier status.

Washington and Beijing debate how fast AI should move, India must ask where

On Nov.15, 2026, the United States and nine other participants, including India, signed the San Francisco Declaration under the Pax Silica initiative. The resolution aimed to secure the supply chain and reduce external reliance on critical components such as semiconductors, crucial to the artificial intelligence infrastructure. 

While the global race for AI has intensified, countries remain divided on how fast it should advance. In Washington, two opposing perspectives have emerged; the decelerationists who seek a slow, cautious and a steady progress and, accelerationists that have found their strongest support in President Donald Trump who has doubled down, by proposing an AI Force, modeled after the U.S. Space Force. 

China, the country driving much of America’s anxiousness has moved ahead through its largely state directed approach, the AI Plus initiative of horizontal integration across sectors. As its newer models close the capability gap Beijing has kept its R&D engines running, refusing to regulate at the expense to innovate. In a rather one-size fits all trajectory, the discourse has therefore been reduced to who must reach the frontier first without deliberating on the speed, based on the real-world applications of AI. 

For a diverse and developing economy like India, that also signed the declaration and is playing catch-up with creating its own indigenous frontier model, the questions posed must be more distinct to its own circumstances. These include: which sectors AI must be readily incorporated?, how fast should the integration process be? and, the extent should it be customized depending on specific needs? 

India’s initial engagement with the domain comes from a position of strength with a large pool of talent to acquire from, a strong consumer market that is expected to reach $131 billion by 2032 and a limited but sector specific models for application. For many however, the destination for success in the field is tied to the number of frontier models owned and how broadly they are being used by other countries, and on such grounds India’s recognition as a leading force is yet to materialize. 

Importantly, India’s stance towards AI adoption even among this upheaval has remained consistent, with questions of moderation emerging in countries that have reached the frontier end. In India, whose integration at present are limited and oriented more towards domestic needs, the debates have instead reinvigorated the diffusion process with New Delhi realizing the ground it must cover with adequate guardrails to rein in the “mother of all tech”. 

Debate, Develop and Deploy

For many therefore when viewed through the lens of the speed of deployment,  AI in India is largely a settled debate. It however took China almost two years to develop and release its DeepSeek R1 model behind ChatGPT and, it will take even longer as it continues to build an entirely self-reliant hardware ecosystem from data centers to semiconductor industries. Other than the U.S. and China, countries such as France with its Mistral AI, Japan’s Sakana AI, and Canada’s Cohere also find themselves approaching the frontier milestone. 

With India’s goals of employing advanced models towards socio-economic inclusion, reproducing the same energy and capital intensive path as those at the forefront of the race, is therefore neither necessary nor feasible. The desire for membership of an exclusive club of frontier models should also not be at the expense of its determined objectives. 

For India the accel/decel debate has a deeper meaning, and its roadmap for execution must be more nuanced consisting of a mix of approaches. This can be done by dividing the areas for intervention and pursuing them simultaneously while managing the speed of deployment to ensure that the objectives and ambitions anchored in sovereign, scale models are achieved more swiftly. 

Firstly, a baseline level of adoption should be ensured where AI handles routine primary tasks with low-stakes such as language and assistance to public services. Under the AI for All endeavor this section has witnessed a considerable focus, through translation platforms such as BHASHINI to targeted language models such as Adi Vaani for tribals. Moreover AI voice bots are now being integrated at scale from railway booking services, finance to public grievance redressal platforms both as part of convenience and a newfound personalized solution based access. 

Then there are sectors where integration of AI has brought rapid transformation, making it a necessary capability that must be quickly covered through approaches like leapfrogging. This includes agriculture, healthcare, education, disaster preparedness, transport and logistics.  It is here that a significant share of time and resources will be required to optimize the benefits, such as in farming through guidance in crops, weather, subsidies and market price discoveries along with healthcare where telemedicine based AI diagnosis saw 282 million consultations between 2023 to 2025 in a country, where professionals are scarce and rural areas suffer from poor connectivity. 

This has made AI a force-multiplier and for many relied on public infrastructure a symbol of inclusive growth. The recent Nepal floods also demonstrated its use in relief response and its implementation in developing countries has transformed urban public transport, showing that adoption at scale of low-cost but high-impact platforms is necessary to harness the benefits of this growing technology.

Finally, the urgency to deploy should be balanced by selectively exploring the risks and its consequences in cases like defense projects and cybersecurity. Earlier in 2017, President Putin had stated, “the one who becomes the leader in this sphere (AI) will be the ruler of the world.” Subsequently its impact on the battlefield has been witnessed in the Russo-Ukraine war where both sides have increasingly turned to AI powered drones and also in the ongoing Houthi-Saudi conflict for their precision-guided ballistic missiles. In other words, use of AI in military has become a crucial affair, with global military-AI spending expected to reach $38.8 billion by 2028.

 India in its policy document “Artificial Intelligence in Military Domain” has also emphasized its opportunities to command, train and deploy forces by enhancing the capability of its assets. However, despite the seriousness of security matters, faster development cannot be the sole objective as failures in this domain can lead to irreversible impact. The Indian Computer Emergency Response Team (CERT) in the first officially detected incident of rogue AIs, reported them going through confidential data along with the role of open-source models being exploited for financial gains. Research and capability building can therefore continue, but deployment must be selective particularly where the pressure to accelerate ends up ignoring the associated harm.

A safety-alignment gap also exists in India. It has no comprehensive statute for AI regulation, with only an IndiaAI Safety Institute been set up to check the risks and challenges, discussions are underway for stronger guardrails. As India aims to move forward with its AI development, regulating the pace will also depend on the advancements made in its ecosystem. 

At present the capabilities remain structurally limited and largely uneven across the country, such as in cloud data centers whose approximate capacity is 1,280 megawatts with Mumbai and Navi Mumbai making up about 25% of the total capacity followed by Bengaluru and Hyderabad at 22% each. Similarly the concentration of the talent pool is largely visible in the cities of Bengaluru, Hyderabad and Mumbai with the clusters of GenAI startups at  Karnataka (39%) and Maharashtra (14%) making up half of the total startups. India’s AI growth is therefore not at an uniform stage and its deployment must reflect the sectoral needs and structural gaps, as the opportunities and consequences from failure will vary across the chain.

From Host to Participant: A Changing Role

The hosts of the 2026 “AI Impact Summit,” New Delhi will soon join a forum of countries at the upcoming discussion on “Bridging Innovation and Trust,” the theme of the 2027 Geneva AI Summit, this time as a participant. After having the New Delhi Declaration being endorsed by 92 countries, with announcements of $250 billion in investments, the time has come for India to advance this debate. 

Moving beyond the binary of acceleration and deceleration and treating frontier AI as the ultimate model of progress, India can advance the question for the developing countries of the Global South for whom sovereign capability, dependence and affordable access of models are as important as the discussions of safety and risks associated with hallucinating, recursive self-improving AI and rogue agents. A blanket call to slow down ends up excluding these concerns and ignore the stakes that many of these countries possess in governing the technology and shaping their future.

Even with all this clarity to advance growth and Indians being optimistic about AI, the roadway ahead remains rugged. Firstly, funding remains a leading concern as even with domain specific AI deployment the fiscal realities and energy capacity can constrain the ease with which the diffusion process can take place as highlighted by the Stanford AI Index on global private investment for the year 2013 to 2025 with the U.S. ($757.27 billion) leading the chart, followed by China ($131.83 billion) and India ($15.39 billion). Secondly, the technology has to be deployed by stacking it on existing infrastructure, such as for farmers using AI services in rural areas a precondition would be reliable internet connections or for healthcare professionals an initial high capital investment of smart sensors is necessary. Finally, building the facilities of AI infrastructure such as data centers have increasingly come across scrutiny by local residents worried about its environmental impact resulting in campaigns such as “We cannot drink DATA.” 

These constraints show that even with a clear direction the path ahead remains difficult. The countries at the frontier end can afford a “two steps forward and one step back” approach India however needs to align its pace based on sector specific needs, available resources and local sentiments so that it doesn’t need to retrace its steps. Emphasizing on the legacy of the February summit, India can thereby truly pursue its agenda of AI for All where the aim is not simply to benefit from AI, but one where everyone is able to participate at their own pace.

About Soumya Dutta

Soumya Dutta is an early career analyst based in New Delhi. He holds a Masters degree from Jamia Millia Islamia, New Delhi. His research interests lie at the intersection of foreign policy, technology and security affairs


View all posts by Soumya Dutta →



Source link

You may also like

Leave a Comment