India’s Defence Self-Reliance Problem: We Keep Building the Platform Before Solving the Choke Point
India is producing more defence equipment at home than ever before. But beneath the impressive production numbers lies a recurring weakness: the country can increasingly build complex platforms without necessarily controlling the engines, seekers, sensors, materials and electronics that determine whether those platforms can be produced, sustained or upgraded independently.
DEFENCE ANALYSIS
In September 2026, 27 Tejas Mk1A airframes were ready at HAL, yet the programme was still working through engine availability and remaining integration issues. 11 GE F404-IN20 engines had arrived, with more expected by the end of the year. The image is striking: India can manufacture the fighter, but a critical foreign subsystem can still determine how quickly the aircraft becomes an operational asset.
The problem, however, is much bigger than Tejas. India has encountered versions of the same problem for decades. The HF-24 Marut was an Indian-designed fighter whose potential was constrained by its underpowered Orpheus engine. The Arjun tank spent years dealing with imported powerpack issues. The Tejas programme itself was supposed to eventually move away from foreign propulsion through the Kaveri engine, but more than three decades after the programme began, India still relies on imported turbofans for its fighters.
At sea, Indian warships have depended on Ukrainian marine gas turbines. In helicopters, the Shakti engine represents genuine Indian manufacturing capability but remains based on a foreign design. In submarines, Indian shipyards can build sophisticated boats while critical technologies such as air-independent propulsion and combat systems continue to involve foreign partners. In missiles, India has achieved impressive levels of indigenous production while some seekers and propulsion technologies remain dependent on external sources.
The pattern is remarkably consistent: India keeps getting better at building the platform, but it has been much slower at owning the technologies underneath the platform. That distinction matters because a weapons system is only as independent as its most difficult-to-replace critical component. A fighter without an engine is an airframe, a missile without its seeker is a projectile, a warship without propulsion is a hull, and a radar without its semiconductor and transmit-receive ecosystem is little more than an antenna assembly. The headline indigenous-content percentage does not capture that vulnerability, and this is where India's defence-industrial story becomes more complicated than the production and export numbers suggest.
The Metric That Flatters
The headline figures are not fake. Indigenous defence production rose from about ₹46,000 crore in 2014–15 to ₹1.54 lakh crore in 2024–25, and one official readout put 2025–26 at ₹1.78 lakh crore. Exports went from under ₹1,000 crore a decade ago to ₹23,622 crore in 2024–25, with a later claim of ₹38,424 crore. The ministry says roughly 65 percent of equipment is now produced in the country, against an earlier import share of 65–70 percent. Private industry is about a quarter of output, while 92 percent of contracts in 2024–25 went to domestic firms.
Those numbers represent real progress. The problem is that production value and technological sovereignty are not the same measurement.
SIPRI's figures sit beside the factory data without necessarily contradicting it. India was still the world's second-largest importer of major arms in 2021–25, accounting for 8.2 percent of global imports. Imports fell 4 percent compared with the previous five-year period. Russia's share dropped from 70 percent in 2011–15 to 40 percent; France rose to 29 percent and Israel to 15 percent.
SIPRI, however, tracks complete weapons rather than every imported component inside an Indian-built system. That distinction is crucial. A country can assemble thousands of platforms domestically while remaining dependent on foreign suppliers for the components that are hardest to replace. The import dependency does not necessarily disappear; it moves down the supply chain.
From Platform Assembly to Technology Ownership
This is where the meaning of technology transfer becomes important. Hollow ToT is not simply the presence of a foreign component in an Indian weapon. It is licensed production without genuine design authority.
The buyer gets drawings, tooling, manufacturing processes and the right to produce a particular system. It may gain valuable industrial knowledge, but it does not necessarily gain the ability to redesign the core technology. It cannot freely change the hot section of an engine, redesign a seeker, modify critical software, substitute a new alloy and independently certify it, or develop and export a derivative without the original technology owner's involvement.
That difference can disappear inside an indigenous-content percentage. A platform may contain a large share of Indian-made components while the most strategically important technology remains foreign.
The GE F414 arrangement for Tejas Mk2 and the first AMCA batch illustrates the distinction. Technical terms were reported agreed in April 2026, with around 80 percent technology transfer and a HAL assembly line potentially following the contract. That would represent significant manufacturing capability, but 80 percent ToT without the underlying design authority is still fundamentally different from owning the engine.
Safran's competing proposal for a 120 kN engine, developed with GTRE and presented as involving full Indian intellectual property including the hot section, therefore belongs to a different category of technology transfer. As of early September 2026, however, that proposal had not cleared the Cabinet Committee on Security.
The question is not whether India should manufacture foreign-designed systems. It should, where that is the fastest way to build capability. The question is whether manufacturing eventually becomes a bridge to design independence or becomes the destination.
The Propulsion Trap
Propulsion is the clearest example because engines combine metallurgy, thermodynamics, precision manufacturing, controls and decades of accumulated testing knowledge.
The Kaveri programme is more than thirty years old. A dry version reportedly reached 48.5 kN during Russian high-altitude testing, above its 46 kN target, and the technology is now being considered for the Ghatak programme. An afterburning configuration was ground-tested in February 2026 in the 81–83 kN class, with a crewed Tejas testbed hoped for around 2030.
But the immediate requirements are larger. The Tejas Mk1A requires a mature 84 kN-class engine, while AMCA Mk2 is being discussed around the 120 kN class, with future growth toward 140 kN. Kaveri is therefore an important technology reservoir, but it is not yet a replacement for the mature engines India needs today.
The same story appears in helicopters. The ALH Dhruv and Prachand use the Shakti engine, a Safran Ardiden derivative manufactured in India by HAL under licence. For the 13-tonne Indian Multi-Role Helicopter, HAL and Safran signed in August 2026 to co-develop the Aravalli, a 3,500–4,000 shaft-horsepower turboshaft through their SAFHAL joint venture, with design work aimed at 2032–33.
That is a more advanced form of cooperation because HAL gains exposure to major engine modules and development processes, but it also demonstrates the timeline problem: India is still building the technological base for the engines its next generation of heavy helicopters will require.
At sea, the dependency has already demonstrated its strategic consequences. Frontline Indian destroyers have historically relied on Ukrainian Zorya-Mashproekt marine gas turbines. The war in Ukraine disrupted that supply and maintenance ecosystem, exposing a vulnerability that had previously been largely invisible during peacetime. BHEL has reported progress on a 40 MW-class indigenous marine turbine, but until such engines enter reliable series production and can be supported throughout their life cycle without foreign spares, Indian warship construction remains vulnerable at the propulsion layer.
The same logic applies to armoured vehicles. The Arjun's history with imported powerpacks demonstrated that building an indigenous hull and turret does not automatically produce an indigenous tank. The platform is only as independent as the subsystem that can stop it from moving.
The Sensor and Seeker Problem
Sensors and seekers present a somewhat different picture from propulsion. Here, India has made considerable progress and has demonstrated that it can design, develop and increasingly manufacture sophisticated systems domestically. The challenge is less about technological capability and more about building the industrial depth required to produce these systems at scale, continuously upgrade them and control the critical components underneath them.
The Uttam AESA radar is a good example of this transition. Developed by LRDE, the radar has undergone extensive flight testing and is progressing towards integration with Indian fighter platforms. The industrial ecosystem around it is also expanding. Astra Microwave's ₹2,205 crore order from HAL for 122 airborne active antenna arrays and 121 interface frames is an example of Indian private industry moving beyond component supply towards the production of sophisticated radar hardware. This is precisely the kind of ecosystem development that strengthens long-term self-reliance.
The important question is therefore not whether India can build an AESA radar. It clearly can. The deeper question is how much of the technology stack beneath that radar is domestically controlled. Transmit-receive modules, gallium-nitride semiconductor devices, advanced packaging, thermal management, processors and specialised electronic components can determine the performance, upgradeability and supply resilience of the final system. India has developed substantial expertise in radar design and system integration, but the semiconductor and component ecosystem still has considerable room to mature.
The seeker landscape is similarly more encouraging than the broader narrative of dependency might suggest. India has developed and flight-tested indigenous seekers across different missile programmes, including RF seekers for air-to-air missiles, while systems such as Akash and other indigenous surface-to-air missile programmes have demonstrated India's ability to develop domestic guidance and seeker technologies. These achievements show that India is not starting from scratch in this field.
The challenge is taking that capability from individual successful programmes into a broad, high-volume ecosystem. Advanced RF seekers, imaging-infrared seekers, detectors, cryogenic coolers, signal-processing electronics and specialised semiconductor components require precision manufacturing and testing infrastructure that must be available consistently across multiple programmes. A successful seeker trial is therefore an important milestone, but building hundreds or thousands of reliable seekers, supporting them through their service life and rapidly introducing improved generations is a much larger industrial task.
Astra's growing role in radar hardware, along with indigenous seeker development, illustrates an important point about India's defence-industrial evolution: in several sensor technologies, the country has already crossed the basic technology-development barrier. The remaining challenge is to shorten the journey from development to certification, production and widespread operational deployment.
BrahMos provides another useful example. With indigenous content already in the 80-percent-plus range, it remains one of India's strongest examples of progressively increasing localisation in a complex weapon system. At the same time, the remaining foreign-dependent elements matter because they are concentrated in technically important areas such as propulsion and certain subsystems. Increasing indigenous content from roughly 83 percent towards
The Layer Beneath the Subsystem
The deeper India goes into the supply chain, the more difficult the problem becomes.
Rare-earth permanent magnets provide one example. India imported 53,748 tonnes of rare-earth magnets in 2024–25, with more than 90 percent reportedly coming from China. In April 2025, Beijing introduced licensing requirements covering seven heavy rare earths, including dysprosium and terbium, which are important for high-performance permanent magnets.
These materials are not exotic curiosities. They can end up in radar actuators, missile-control systems, electric motors and generators. A fighter assembled in Bengaluru can therefore be affected by a supply-chain decision made thousands of kilometres away.
Specialty materials create another vulnerability. High-strength aluminium-lithium alloys, titanium alloys, carbon fibre, propulsion-grade materials and the single-crystal superalloys required for modern turbine blades are difficult to localise because they require not only raw-material production but specialised processing, certification and repeatable industrial quality.
India can announce a new materials plant in a defence corridor, but that does not immediately create a certified turbine-blade ecosystem.
Electronics: The Quiet Dependency
Electronics are perhaps the least visible layer of the problem. A mission computer can be assembled in India while its processors are foreign. A radar can be designed by an Indian laboratory while its semiconductor fabrication depends on an overseas foundry. A missile can have an Indian guidance algorithm while the underlying detector, processor or high-reliability electronic component is imported.
Radiation-tolerant and high-reliability electronics remain particularly difficult. So does the software infrastructure underneath them. Flight-control laws, electronic-warfare libraries and mission software can increasingly be developed domestically, while operating systems, compilers, electronic-design-automation tools, cryptographic components and other elements of the development ecosystem remain dependent on foreign technology.
This dependency rarely appears in an indigenous-content percentage, but it can nevertheless become strategically decisive. A denied software update can be just as consequential as a denied spare part.
Why the Indigenisation Lists Do Not Solve It
India's positive indigenisation lists are useful. They create predictable opportunities for Indian companies and help MSMEs enter defence supply chains.
The sixth list, notified in August 2026, covers 405 items with an estimated business value of ₹3,070 crore, including line-replaceable units and spares for platforms such as the ALH, Su-30MKI, Tejas, AL-31FP, T-90 and BMP-II.
Across the portal, more than 15,700 items have reportedly been indigenised, representing roughly ₹9,000 crore in import substitution over five years.
That is meaningful industrial development, but it also reveals a structural bias. The easiest things to indigenise are usually the things Indian industry can already manufacture: machined parts, castings, cables, brackets, pumps, fasteners and conventional electronics.
The hardest items remain difficult precisely because they require capabilities India has not yet built at scale. Single-crystal turbine blades, advanced infrared detectors, high-end semiconductor fabrication, advanced inertial sensors, modern afterburning turbofans and high-temperature materials are not notification problems. They are ecosystem problems.
The Problem With Chasing Percentages
This is why another increase in the indigenous-content percentage should not automatically be treated as another increase in strategic autonomy.
Imagine two weapons. The first is 90 percent indigenous, but the remaining 10 percent includes its engine, seeker and critical processor. The second is 75 percent indigenous, but India owns the design authority for every critical subsystem and can manufacture, modify and sustain the weapon without external permission.
Which is more sovereign?
The answer is obvious.
The objective should therefore move from percentage localisation to denial resilience. India needs to identify the relatively small number of technologies whose absence can cripple an otherwise indigenous weapon system: fighter and marine gas turbines, infrared seekers, advanced RF seekers, GaN semiconductor fabrication, inertial sensors, propulsion-grade superalloys, high-reliability processors and safety-critical software.
These are the technologies where the final few percentage points matter more than the first 80.
We Have Seen This Before
The uncomfortable part is that none of this is entirely new. India has repeatedly built increasingly sophisticated platforms while discovering that the hardest technology was somewhere underneath them.
The Marut exposed the propulsion problem. Arjun exposed the powerpack problem. Kaveri exposed how difficult it is to develop a modern fighter engine. The disruption of Ukrainian marine turbines exposed the naval propulsion problem. Tejas has again demonstrated how a foreign engine can become a production bottleneck.
None of these programmes was a failure simply because they depended on foreign technology. In many cases, foreign cooperation or imports were necessary to deliver capability. The failure would be to encounter the same dependency repeatedly without building the ecosystem needed to eliminate it.
That is the real meaning of learning from experience.
The Way Forward
India does not need to stop importing. It needs to become much more selective about what it imports permanently.
Buying a foreign aircraft to fill an immediate capability gap can make sense. Buying a foreign aircraft while simultaneously ensuring that its critical technologies never become indigenous is a different strategic decision. The same principle applies to helicopters, submarines, radars, missiles and armoured vehicles.
The government should identify denial-critical technologies and fund them as long-term national programmes rather than treating them as components to be localised after the platform has already been ordered. Test facilities need to exist before the production requirement arrives. Industry needs guaranteed development pathways rather than sporadic orders. Universities need to participate in materials, propulsion, semiconductor and sensor research. Most importantly, procurement needs to recognise that the difficult technology may take a decade to mature.
That decade cannot begin after the platform enters production.
Operation Sindoor demonstrated how much India's indigenous defence ecosystem has already changed. BrahMos, Akash, MRSAM, domestic drones, indigenous radars, artillery systems and warships are evidence that India is no longer merely assembling foreign weapons. That is a major achievement.
But self-reliance is not achieved when the platform carries an Indian flag. It is achieved when an adversary, supplier or geopolitical crisis cannot remove one critical foreign component and leave the platform stranded.
India has spent decades learning how difficult that is. The next decade should be about proving that we have finally learned the lesson.
Assembly fills the order book. Critical technology determines whether that order book can go to war.