Centre may steer defence and aerospace towards Indian AI models and away from foreign frontier systems
The Centre is reportedly likely to push defence, aerospace and other critical departments to prefer Indian-built AI models such as those from Sarvam and BharatGen, and possibly to avoid foreign frontier models altogether.
India's government is reportedly preparing to put its weight behind home-grown artificial intelligence in the places where dependence matters most. According to a Business Standard report published on 9 November 2026, the Centre is likely to push critical sectors — starting with defence and aerospace — to favour indigenous AI models, and could go as far as asking critical government departments at both central and state level to avoid frontier models and AI agents developed by non-Indian companies.
What is being proposed
The report, citing officials, says the push would cover Indian large language models, sector-specific models and AI agents. Developers such as Sarvam and BharatGen, which have already released several models, were named in the coverage.
Business Standard describes discussions with officials. A formal policy, scope and timetable were not established by that reporting.
Why the government is moving in this direction
The logic is about control. Some frontier AI systems rely on overseas hosting and vendor-managed updates, and their behaviour, availability and terms of use can change without notice. For defence planning, aerospace engineering or sensitive administration, that creates risks around data leaving the country, supply interruptions and opaque model behaviour.
The defence establishment has been saying this publicly. At the India-AI Impact Summit 2026, DRDO's Director General said that in the defence domain, India "can't afford to depend on solutions and AI models which are coming from abroad." And at the Delhi Defence Dialogue on 3 November, Defence Minister Rajnath Singh urged that self-reliance extend beyond hardware to control over the algorithms, data and chips that power military platforms — including secure indigenous software stacks, trusted semiconductor supply chains and home-grown AI models.
The domestic model pipeline
The policy would lean on capacity built under the IndiaAI Mission, which has an outlay of ₹10,372 crore and under which the government has approved the development of 20 AI models. On 3 November 2026, the mission released three more indigenous foundation models in New Delhi: EKA-1 from Soket AI Labs and Dhīr 2B and Dhīr 5B from ZenteiQ AiTech Innovations.
A preference mandate could give these developers something they have struggled to find: large, steady government demand. For sovereign-model builders, procurement by ministries, armed forces and PSUs would provide revenue, real-world feedback and domain data to fine-tune on.
What it would mean in practice
If formalised, technology leaders in defence, aerospace and other critical public bodies would need to map where foreign frontier models are used today — in coding assistants, document summarisation, translation, analytics or agent workflows — and plan migration paths. That raises practical questions the policy will need to answer:
- Scope: Does "avoid" apply to all use or only to sensitive data and decision-making systems?
- Deployment: Will Indian models run on-premise or in government-controlled clouds, and on whose GPUs?
- Evaluation: How will capability and safety be tested for each workload and deployment environment?
- Open weights: Would open-weight foreign models hosted domestically count as acceptable?
The bigger picture
The move fits a wider pattern of "stack sovereignty" in Indian tech policy, from the semiconductor mission to indigenous processors and public digital infrastructure. Done well, it could anchor a domestic AI industry with serious, mission-critical customers. Done bluntly, it could leave critical users with weaker tools. The details — expected in any formal guidance — will decide which.
Sources
- https://www.business-standard.com/technology/artificial-intelligence/govt-likely-to-push-for-use-of-domestic-ai-models-in-critical-sectors-126110901119_1.html
- https://www.angelone.in/news/economy/government-pushes-for-greater-use-of-indigenous-ai-models-in-critical-sectors
- https://cio.eletsonline.com/artificial-intelligence/indigenous-ai-models-critical-sectors/76824/
- https://raksha-anirveda.com/cannot-afford-to-depend-on-ai-models-coming-from-abroad-drdo/
- https://ddindia.co.in/2026/11/rajnath-singh-calls-for-indigenous-production-of-critical-defence-technologies-independent-sustenance-of-capabilities/
- https://www.tribuneindia.com/news/delhi/digital-sovereignty-is-also-needed-for-self-reliance-says-rajnath