India’s armed forces are moving to integrate artificial intelligence more deeply into military operations through a five-year agreement between Headquarters Integrated Defence Staff (HQ IDS) and Bangalore-based Sarvam AI.

The partnership aims to create a common AI framework spanning the Army, Navy and Air Force. Rather than developing separate AI tools within individual services, the initiative seeks to establish a shared architecture capable of supporting intelligence analysis, target acquisition, decision support and data fusion across military operations.

A key objective is to improve how information flows through the sensor-to-shooter loop. Data gathered from satellites, radar systems, drones and other intelligence assets can be consolidated and analysed more efficiently, enabling commanders to gain a clearer picture of the battlefield and make faster, better-informed decisions.

The project reflects a broader shift from isolated AI deployments towards a connected military ecosystem. By linking datasets, platforms and operational processes, the armed forces hope to improve situational awareness and accelerate operational decision-making.

Sarvam AI was selected partly because of its sovereign AI approach. The company develops its own foundation models and AI infrastructure within India, allowing sensitive military data, operational plans and classified intelligence to remain under domestic control.

Military planners view this as strategically important because dependence on foreign AI infrastructure could create vulnerabilities involving data security, technology access and operational autonomy.

However, Sarvam is not the only Indian company pursuing indigenous AI. The agreement does not imply exclusive capability. Industry observers note that Sarvam’s ability to combine foundation models, agentic infrastructure, geospatial analytics and perception systems within a unified framework may have made it particularly suited to the armed forces’ cross-service ambitions.

Under the agreement, Sarvam will design, develop and deploy AI applications for operational requirements, training and decision support. The company will also provide technical support after deployment. Financial details, specific project plans and implementation timelines have not been disclosed.

The work will be conducted through Chanakya, Sarvam’s dedicated and firewalled national security division. Its technology stack includes the Sarvam 105B foundation model together with agentic systems, geospatial tools and perception capabilities.

Sarvam describes the vision as an “AI Fabric” connecting different elements of India’s defence infrastructure. The intention is for lessons learned in one service to be shared across others, reducing duplication and improving interoperability.

Among the most closely watched aspects of the initiative is its potential role in creating an indigenous AI-enabled kill chain. Such systems can help process large volumes of intelligence, identify objects of interest, correlate information from multiple sources and recommend priorities to commanders.

AI could assist with target identification, mission planning and weapon selection while reducing the time between detection and response. One of the primary goals of military AI worldwide is to compress the kill chain and accelerate operational tempo.

Importantly, Sarvam has repeatedly emphasised that human oversight will remain central. The company stated: “The focus areas are decision support, target acquisition and data fusion across the sensor-to-shooter loop, with human oversight retained on operational decisions.”

This means AI systems may recommend targets or courses of action, but authority for military engagement would remain with human commanders rather than autonomous algorithms.

International experience illustrates both the potential and the risks of such technology. Reports concerning the US military’s use of AI-assisted targeting systems during Operation Epic Fury demonstrated how AI can rapidly analyse intelligence from multiple sources and help prioritise targets.

These systems reportedly accelerated processes that traditionally required large teams of analysts to sift through extensive intelligence data before generating target lists and operational recommendations.

At the same time, controversy surrounding civilian casualties during the campaign highlighted concerns about flawed intelligence, outdated imagery and excessive reliance on AI-supported assessments. Such incidents have strengthened arguments for maintaining meaningful human control over military decision-making.

India has already accumulated significant operational experience with military AI. During Operation Sindoor in May 2025, the Indian Army employed AI for intelligence analysis, threat assessment, predictive modelling, weather forecasting and creation of a common operational picture.

Lieutenant General Rajiv Kumar Sahni stated that AI was used extensively for real-time multi-sensor and multi-source data fusion. According to him, 23 specialised applications processed operational data and intelligence inputs. He further said AI-assisted analysis contributed to 94% targeting accuracy against Pakistani military positions and terror infrastructure.

Since then, the Army has demonstrated AI-powered satellite imagery analysis systems capable of detecting changes and highlighting significant observations. It has also showcased portable AI platforms designed to support planning and analysis in environments without internet connectivity.

Another notable initiative is Ekam AI, an indigenous platform developed to process information and support decision-making within secure operational environments.

The Navy has likewise expanded its adoption of AI technologies. Its 2024-25 annual report highlighted Antarbashi.AI, an offline language translation system used during anti-piracy missions, and Satyapan.AI, a facial-recognition solution designed to assist in identifying suspected pirates.

Other naval programs include NARAD, a private offline language model used for searching and summarising official documents, and HawkSecure, which applies AI and machine learning to monitor network traffic and detect anomalies.

The Navy has also tested Drishti.AI aboard INS Visakhapatnam. The system is designed to monitor personnel movement and detect man-overboard incidents. These projects vary in maturity, with some operationally deployed and others still undergoing trials.

The Army is additionally exploring autonomous military systems. A Ministry of Defence report noted that an autonomous weapon system experiment under Mission SWAYAT was conducted at the Babina Field Firing Range in April 2025.

The exercise was intended to evaluate indigenous technologies and develop a roadmap for potential future induction. However, it does not indicate that autonomous lethal weapons have entered operational service.

The Sarvam agreement could become the connective layer that unifies these currently separate efforts. Success will depend on the ability of different systems to work together effectively, protect sensitive information and provide recommendations that remain reliable under combat conditions.

For now, the memorandum of understanding represents a developmental framework rather than proof that India is building a fully integrated AI-enabled kill chain. Its broader significance lies in establishing a path towards sovereign military AI, bringing advanced automation into the intelligence-to-decision process while ensuring that responsibility for battlefield actions remains firmly in human hands.

Agencies