This edition covers three developments. First, a feature on AI in warfare, prompted by the ongoing conflict in the Middle East and a striking confrontation between the U.S. government and one of its own AI firms, and what these episodes reveal about two very different models of integrating AI into national defence. Second, the release of India’s Defence Forces Vision 2047. Third, the 18th India-US Defence Policy Group meeting, held as the West Asia conflict enters its second month.
AI in Warfare
Five weeks into Operation Epic Fury, the U.S.-Israeli campaign against Iran has become the most AI-intensive military operation in history. CENTCOM Commander Admiral Brad Cooper confirmed in March that American forces are deploying a range of advanced AI tools to compress targeting workflows that once required hours, into seconds. Iran, meanwhile, has fired thousands of low-cost drones across the Persian Gulf. The Shahed series, at roughly $20,000 to $50,000 per unit, has struck military installations, oil infrastructure, and civilian targets across a dozen countries. During the first week of Iranian retaliation, drones accounted for roughly 71 percent of documented strikes on Gulf states. What defence planners spent years discussing in the abstract is now unfolding in real time: AI-enabled systems are accelerating the kill chain, the cycle from detection to engagement, while cheap autonomous platforms are rewriting the economics of warfare. A $50,000 drone can disable an oil tanker or punch through an operations centre, which is orders of magnitude cheaper than the $3-4 Million Patriot interceptor that’s being used against these.
The technology is also exposing the costs of speed without adequate safeguards. The Pentagon is investigating whether its AI-assisted targeting system, Project Maven, built on Palantir's platform and incorporating Anthropic's model Claude, contributed to the strike on an Iranian girls' school that killed over 100 people, mostly children. Separately, researchers found that when placed in simulated conflict scenarios, AI models from leading developers signaled nuclear weapons use in 95 percent of cases.
This is the backdrop against which a parallel confrontation played out between Anthropic and the U.S. Department of War over the terms on which frontier AI may be embedded in military operations. In our opening edition, we observed that AI and autonomous systems have the potential to reshape procurement, manufacturing ambitions, and foreign technology partnerships. That observation now reads less like a forecast and more like a description of the present.
The dispute had been building since January, when reports emerged that Claude had been used during Operation Absolute Resolve in Venezuela, a military action Anthropic's own terms of use would prohibit. By late February, the friction turned public. Defence Secretary Pete Hegseth reportedly gave CEO Dario Amodei a deadline: allow unrestricted use for all lawful purposes, or lose access. Anthropic refused. Within a day, Trump directed all federal agencies to stop using its products and Hegseth designated the firm a "supply chain risk," a classification previously reserved for companies linked to foreign adversaries. OpenAI stepped in within hours, accepting the Pentagon's terms. The backlash was significant, prompting subscriber cancellations and senior resignations. Altman later amended the contract to add surveillance prohibitions. Anthropic sued on 9 March. On 26 March, U.S. District Judge Rita Lin granted a preliminary injunction, finding that the designation was likely retaliatory and unconstitutional, prompted not by any substantiated security concern but by Anthropic's public stance. The episode crystallises a structural question that any country integrating frontier AI into military systems will need to answer: when the model is built by a private firm, who draws the line on its use? The American answer, for now, remains contested.
India’s answer begins from a different premise: Rather than contracting with external frontier firms and then negotiating boundaries of use, India has moved toward building sovereign compute and model development infrastructure within national boundaries. The IndiaAI Mission, with an outlay of ₹10,372 crore, has onboarded over 38,000 GPUs at subsidised rates, with a 100% compute subsidy for foundational model builders. Sarvam AI, tasked with building India's sovereign large language model, launched Chanakya on 29 March, an air-gapped, on-premise vertical explicitly designed for dual-use deployment across enterprise and defence. The Indian Army separately established a sovereign AI-ML lab with CoRover in February, built around its BharatGPT platform.
The operational picture is already considerable. Over 75 AI projects are active across services. The Indian Navy has deployed tools ranging from Antarbhashi.AI for multilingual translation in operational environments to TRIDENT-SAMUDRA, an AI-powered ocean surveillance system unveiled in March. CDS General Anil Chauhan stated at the Raisina Dialogue that AI will be central to future warfare and that dedicated energy infrastructure will be needed to support it. The ambitions go further. The Parliamentary Standing Committee on Communications and IT, in its report on AI, confirmed that DRDO and defence public sector undertakings are actively researching lethal autonomous weapons systems, alongside AI applications spanning command-and-control, autonomous underwater vehicles with target detection and classification capabilities, AI-powered missile systems, and border security. The same report noted India's strategic dependence on imported GPU accelerators and flagged large language models for strategic applications as a national need, distinct from civilian use cases, requiring investment in classified datasets held across government departments.
These developments point to an institutional design that seems to largely sidestep the problem the US has been grappling with. By channelling public subsidies into domestic compute infrastructure and indigenous model development, and by building AI tools for defence within sovereign, air-gapped environments, India has structured its AI-defence relationship in a way that keeps the state as both the principal funder and co-developer as well as the primary user. While this does not eliminate tensions around ethical boundaries of military AI, it does mean the question of whether a private vendor can refuse a sovereign's demand is less likely to present itself in the same form. Whether this approach produces models that match the capability frontier of their American counterparts remains an open question. But for a country with two adversarial neighbours and active border challenges, even if the state prefers to have end-to-end control over the AI systems embedded in its military infrastructure, this is not difficult to understand.
In Focus:
Defence Forces Vision 2047
On 10 March, Defence Minister Rajnath Singh released 'Defence Forces Vision 2047: A Roadmap for a Future-Ready Indian Military', a document prepared by Headquarters Integrated Defence Staff outlining strategic reforms and capability goals for the armed forces by the centenary of India's independence. The document is framed as a guideline rather than a directive - several of its goals are acknowledged to be aspirational and contingent on approvals at multiple levels, including the Cabinet Committee on Security.
The roadmap also touches upon the current nature of conflict, arguing that the distinction between war and peace has eroded, that adversaries will increasingly operate through grey-zone coercion and cognitive operations, and that warfare is evolving from network-centric to data-centric and ultimately intelligence-centric models, driven by AI, autonomous systems, and edge computing. The institutional proposals that follow this diagnosis are specific: the creation of a Data Force, a Drone Force, a Cognitive Warfare Action Force, and a Defence Geo-Spatial Agency, alongside dedicated Space and Cyber Commands. The document also outlines Mission Sudarshan Chakra, an expansion of ballistic missile and air defence systems to protect strategic and civilian infrastructure against multi-domain threats.
On jointness, the document calls for establishment of a Joint Headquarters and operationalising a Joint Operations Coordination Centre. The three-phase roadmap is structured accordingly: a transition phase through 2030, focused on multi-domain restructuring and net-centric operations; a consolidation phase through 2040, centred on indigenous capability development for data-centric warfare; and an era of excellence by 2047, in which the forces are envisaged to possess decision superiority across all domains.
The document has since drawn both support for its conceptual clarity and criticism for restating goals - jointness, self-reliance, technological modernisation that have featured in successive reform exercises over two decades with implementation falling short according to critics.
18th India-US Defence Policy Group (DPG) Meeting
On 25 March, India and the United States held the 18th Defence Policy Group meeting in New Delhi, co-chaired by Defence Secretary Rajesh Kumar Singh and US Under Secretary of War for Policy Elbridge Colby. The DPG is the apex official-level mechanism between the two defence establishments, and this meeting was the first to take place under the 10-year Framework for the US-India Major Defence Partnership signed by Defence Minister Rajnath Singh and US Secretary of War Pete Hegseth in Kuala Lumpur in October 2025.
The Ministry of Defence stated that both sides reviewed ongoing initiatives, identified priority areas for co-development and co-production of defence equipment, and reaffirmed commitment to enhancing military-to-military cooperation through joint exercises, training visits, and strategic exchanges. The discussions reportedly covered specific procurement items, including Excalibur precision-guided munitions under a ₹300 crore deal already signed, potential acquisition of Javelin anti-tank guided missiles through the emergency route, and India's plans for procuring six additional P-8I maritime surveillance aircraft from Boeing.
A day before the DPG meeting, Colby delivered a speech at the Ananta Centre in New Delhi that went beyond customary diplomatic practice. He described India as "indispensable" to maintaining a stable balance of power in the Indo-Pacific and positioned the partnership in explicitly realistic terms drawing on External Affairs Minister Jaishankar's own articulation of a hard-headed, interest-driven foreign policy. He was direct in stating that the United States wants partnerships with "vigorous, self-assured states, not dependencies" and that disagreement between the two countries need not impede strategic cooperation where interests converge. On defence industry, he acknowledged India's ambition to expand its indigenous defence base as "entirely reasonable" and committed to advancing co-production and co-development in areas including long-range precision fires, anti-submarine warfare, maritime domain awareness, and advanced technologies.
Colby's Ananta Centre address in particular seems to represent a conscious effort to frame the defence relationship not as a mere transactional buyer-seller arrangement but as a structural alignment between two states whose interests independently converge on regional stability. Whether the institutional machinery on both sides - regulatory barriers, procurement differences, and bureaucratic inertia, which Colby himself acknowledged as "real challenges" can deliver on that framing will determine whether this decade of the Major Defence Partnership advances beyond its predecessors.
Defence Roundup:
- The Ministry of Defence signed contracts worth ₹5,083 crore for the procurement of ALH Mk-III helicopters from Hindustan Aeronautics Limited for the Indian Coast Guard, and VL-Shtil surface-to-air missiles from JSC Rosoboronexport for the Indian Navy. The missile deal, signed amid the backdrop of U.S. sanctions on Russia, is intended to enhance warship air defence capabilities.
- Mazagon Dock Shipbuilders confirmed the completion of Contract Negotiation Committee talks for a ₹99,000 crore defence deal with the Indian Navy, with the proposal now awaiting approval from competent government authorities.
- Solar Defence and Aerospace announced an investment of over ₹12,800 crore to expand manufacturing of UAVs, robotics, and long-range missile systems, including development of MALE-class unmanned aircraft and plans to build what it described as India's first AI-powered Industry 5.0 defence manufacturing ecosystem.
- Indonesia reportedly reached an agreement with India to procure the BrahMos supersonic cruise missile system, in a deal estimated at USD 200–350 million. Separately, India briefed Vietnam and Indonesia on the Astra Mk1 beyond-visual-range air-to-air missile as part of its expanding defence export outreach in Southeast Asia.
- The Defence Procurement Board recommended procurement of 60 units of DRDO's indigenous Ghatak stealth unmanned combat aerial vehicle, a flying-wing platform designed for deep-strike missions and suppression of enemy air defences, pending final approval from the Defence Acquisition Council.
- The Defence Acquisition Council granted Acceptance of Necessity for ₹2.38 lakh crore in procurement proposals across the Army, Air Force, and Coast Guard, covering air defence, artillery, communications, surveillance, and aviation. Key approvals include S-400 systems, medium transport aircraft, remotely piloted strike aircraft, Su-30 engine upgrades, Dhanush artillery, and air cushion vehicles for coastal operations.
- India and Canada launched a formal India-Canada Defence Dialogue during Prime Minister Mark Carney's visit to New Delhi, covering defence industries, maritime domain awareness, and military exchanges, alongside agreement on a new Maritime Security Partnership.
Defence Time Machine:
Long before medieval siege engines became standard, the kingdom of Magadha under King Ajatashatru (c. 5th century BCE) is traditionally credited with deploying some of the earliest named mechanical weapons in Indian warfare. Ancient sources describe the mahashilakantaka - a large stone-throwing engine used in sieges and the rathamusala, a war chariot fitted with protruding blades or maces designed to break infantry formations. While partially preserved through later textual traditions, these references indicate that early historic India was already conceptualising and naming battlefield machines, not just relying on conventional infantry or cavalry combat.
For any queries reach out to us at contact@ikigailaw.com
