General David Petraeus’s warning at the NDTV World Summit carries significant implications for India’s future military planning. His central argument was that armed forces worldwide can no longer rely on gradual modernisation cycles.

With artificial intelligence, autonomous systems and emerging technologies evolving in real time, militaries must be prepared to transform concepts of warfare, training systems, procurement processes and organisational structures within a year rather than over traditional five-year timelines, The Print reported.

India undoubtedly faces the same urgency. However, the country must avoid blindly replicating lessons drawn from foreign conflicts or adopting technological models developed for vastly different strategic environments. The more important challenge is determining which technologies suit India’s unique requirements and how they should be integrated into national defence planning.

Recent conflicts in Ukraine and the Middle East have demonstrated that technologically innovative and determined actors can significantly erode the advantages traditionally enjoyed by larger military powers. Indigenous solutions, tailored to local operational realities and supported by emerging technologies, have repeatedly disrupted conventional assumptions regarding military superiority.

For India, the lesson is not simply to acquire advanced technology. The real lesson is to combine technological sovereignty with military doctrines designed specifically for India’s operational environment. Long-term deterrence requires defence capabilities that are developed, sustained and improved domestically rather than through dependence on foreign suppliers or off-the-shelf systems.

India’s requirements differ fundamentally from those of the United States. American military planning is heavily influenced by expeditionary operations, long-range force projection and global intervention capabilities. India’s strategic priorities are centred on defending contested borders, deterring aggression, maintaining regional stability, protecting its maritime interests and ensuring rapid response options in times of crisis.

As a result, India’s military AI architecture must focus on persistent surveillance, rapid decision-making, territorial defence and border management rather than replicating models designed for distant interventions.

Geography further reinforces this requirement. India’s operating environment spans high-altitude Himalayan regions, deserts, jungles, plains and extensive maritime zones. Equipment that performs effectively elsewhere cannot automatically be expected to succeed under Indian conditions. Drones, sensors, communications systems and autonomous platforms must all be rigorously tested in real deployment environments before entering service.

Affordability will remain equally important. India maintains a large standing military force with extensive operational commitments. Any transformation effort must therefore balance innovation with cost, reliability and maintainability. Indigenous production, repair and upgrade capabilities are critical for reducing vulnerability to foreign pressure during periods of conflict.

The transformation required is not solely technological. It represents a fundamental cultural shift. Military leaders, institutions and training establishments must develop a far deeper understanding of algorithm-driven warfare and data-centric operations. AI cannot simply be purchased and integrated like a conventional weapons platform.

The article argues that military AI should be built through a ground-up indigenous effort involving users, developers, researchers and defence institutions. Ownership of this transformation must rest with the armed forces themselves.

A central theme is the importance of sovereign AI. India’s military future, according to the analysis, cannot be built on foreign frontier models, overseas cloud services or open internet dependencies. Instead, India should prioritise the development of indigenous foundational models capable of operating independently within secure national ecosystems.

Sensitive defence information must remain within protected domestic infrastructure. The article highlights secure indigenous vector database solutions such as ShaktiDB, developed through IITM Pravartak, as examples of how operational data can remain insulated from external access.

India also possesses a major competitive advantage in the form of its extensive repository of operational experience. Decades of combat exposure, border management, counter-insurgency operations and international peacekeeping missions have generated a vast pool of real-world military knowledge spanning glaciers, deserts, jungles, riverine environments and littoral regions.

Transforming this experience into structured, high-quality datasets could provide the foundation for highly specialised military AI models tailored specifically to Indian operational requirements. Such systems would be difficult for any foreign competitor to replicate.

The article also highlights the importance of India’s deep-tech ecosystem. IITs, research institutions and innovative Start-Ups already possess significant technical expertise. The challenge is not a lack of talent but the creation of institutional mechanisms that allow these organisations to contribute meaningfully to national security requirements.

Another key argument concerns the future of language models in defence applications. The article contends that massive cloud-based Large Language Models are often expensive, infrastructure-intensive and potentially problematic from a data sovereignty perspective.

Instead, it advocates greater emphasis on Small Language Models deployed directly at the edge. These systems can be optimised for specific operational tasks, offer lower computational overheads and operate in environments where connectivity, security and real-time performance are critical.

Rather than pursuing a single all-encompassing AI platform, the article proposes a modular architecture similar to a Mixture of Experts approach. Under such a framework, specialised models responsible for text analysis, imagery, signals intelligence and other functions would collaborate to generate comprehensive situational awareness.

This approach would allow faster updates, lower costs and greater opportunities for innovation from domestic developers while maintaining flexibility across multiple operational domains.

The analysis also challenges the view that emerging drone threats make traditional military platforms obsolete. Vulnerability does not automatically equal irrelevance. Tanks, ships, infantry formations and other conventional assets continue to perform essential battlefield roles when integrated within modern multidomain forces.

As the article notes, warfare is not a contest between drones and tanks. Effective military power arises from combining mobility, protection, sensors, electronic warfare, communications and counter-drone systems into a coherent operational structure.

"Drones alone cannot hold territory, and tanks cannot survive without electronic warfare and counter-drone defences."

The broader conclusion is that India should use the current period to build, test and refine future capabilities before conflict forces emergency adaptation. Accelerated procurement, extensive testing, realistic exercises, secure communications, resilient cyber capabilities and intelligent logistics must all form part of a larger effort to create an indigenous and self-sustaining model of military modernisation.

Ultimately, India’s military AI transformation, the article argues, should not be imported. It must be built around sovereign technology, indigenous innovation, operational experience and strategic realities unique to India. Only then can the country achieve genuine technological independence and a durable deterrent suited to the demands of modern warfare.

Agencies