Madhav Sethi
Geopolitics · 24 min read

Genuine Reform, Real Contradiction, Fixable: India's DAP 2026, Operation Sindoor, and the Tempo Problem

Whether India's new defence-buying rulebook — DAP 2026 — answers the war it just fought: genuine reform, a real internal contradiction, and the fix.

By

The short version

  • DAP 2026 is genuine, serious reform with a real internal contradiction. The draft — the Defence Acquisition Procedure, India’s official rulebook for buying military equipment — was released for comment on 10–11 February 2026. It pushes speed (technology-readiness-based categorisation, a refined Fast Track route, the iDEX startup channel and the “Make” route elevated to mainstream pathways with five years of assured orders, a new Low-Cost Capital Acquisition route) and, simultaneously, deeper sovereignty (an “Owned by India” intellectual-property and source-code mandate, indigenous content raised from 50% to 60%, and the exclusion of foreign-owned Indian subsidiaries from “Indian vendor” status). The tension is that the heavy indigenisation machinery built for big platforms collides with the software-speed iteration that fast-moving categories — drones, counter-drone systems, electronic warfare, loitering munitions (cheap one-way “kamikaze” drones) — actually require.
  • Operation Sindoor (7–10 May 2025) revealed that the binding constraint is TEMPO, not ambition. The things that worked at speed came through the Emergency Procurement route that bypasses the normal DAP system — 13 contracts worth ₹1,981.90 crore (~$232 million) — and even “indigenous” systems depend on imported batteries, motors, sensors and chips. The Defence Secretary himself called Sindoor a “reality check” exposing gaps in electronic warfare, counter-drone systems and military-grade drones.
  • The contradiction is fixable with a carve-out. Ringfence the fast, iterative, software-speed categories from the 60% indigenous-content floor and the documentation-heavy “indigenous design” tests built for fighters and warships — while keeping the sovereignty mandate where it belongs (platforms, source code, lifecycle control). Net verdict: not triumphalist boosterism, not stale cynicism — a serious reform that needs one structural fix to avoid preparing for the last war.

What stands out

1. DAP 2026 is still a DRAFT, not notified law. The Ministry of Defence (MoD) released the draft on 10–11 February 2026, inviting stakeholder comments by 3 March 2026. It is an ~800-page, two-volume document (the first procurement procedure, in 2002, ran to 84 pages); a procedure for aerospace systems is to be added later. It is intended to take effect from 1 April 2026, synchronised with the FY27 capital procurement budget of ~₹2.19 lakh crore. As of writing it has not been finalised or notified.

2. The core changes from DAP 2020. Categories cut from five to four (the standalone “Buy (Indian)” route dropped); indigenous content raised from 50% to 60% in the top “Buy (Indian-IDDM)” category (IDDM = Indigenously Designed, Developed and Manufactured); even “Buy (Global)” now carries a 30% indigenous-content requirement (previously zero); “indigenous design” defined for the first time around ownership of design documents, source code, system architecture and critical technical data; technology-readiness-based categorisation; a refined Fast Track Procedure with decisions pushed to lower levels; two-stage trials; the iDEX/Make innovation routes given five years of assured orders and “spiral” (iterative) development; and new Long-Term Bulk Acquisition and Low-Cost Capital Acquisition pathways.

3. The Emergency Procurement route — not the mainstream DAP — delivered Sindoor’s tempo. 13 Emergency Procurement contracts worth ₹1,981.90 crore covering counter-drone systems, radars, very-short-range air defence, loitering munitions (including ~450 Nagastra-1R from Solar Defence & Aerospace), drones and soldier protection. Services were authorised to spend up to 15% of their capital budgets on Emergency Procurement. That the things that worked came through the bypass route is itself the key signal about the mainstream system.

4. Indigenous systems performed — but ride on imported guts. The Akash and Akashteer air-defence systems and the integrated counter-drone grid performed well. But India’s drone makers import batteries, motors, rare-earth magnets, sensors and flight-controller chips, mostly from China. In high-altitude jamming trials near Dehradun in late September 2025, every one of 46 indigenous drone manufacturers summoned reportedly failed to operate reliably under simulated jamming, spoofing and signal-hijacking. A deliberately harsher October re-test — using the Army’s full counter-drone arsenal — was passed by only a handful of firms (Raphe mPhibr, NewSpace Research & Technologies, SMPP and Munitions India Limited). The episode prompted a ₹500 crore (~$53 million) National Military Drone Technology Hub at IIT Kanpur, to build jam-resistant, frequency-hopping communications that do not depend on foreign chips.

5. The numbers confirm a sovereignty bet at scale. FY2026-27 defence budget ₹7.85 lakh crore (highest ever, +15.19%); capital ₹2.19 lakh crore (+21.84%); 75% of capital acquisition (₹1.39 lakh crore) reserved for domestic industry. But pensions (₹1.71 lakh crore) plus pay still dominate the total. Record capital contracts of ₹1.82 lakh crore signed in FY2025-26 (to December). Defence production ₹1.78 lakh crore; exports a record ₹38,424 crore in FY2025-26 (+62.66% on the prior year’s ₹23,622 crore).

6. The 114-Rafale deal embodies — and is now stuck on — the tension. In-principle approval came in February 2026 for ~₹3.25 lakh crore / ~$40 billion (18 flyaway aircraft plus 96 made in India, F4 standard, a ≥60% indigenous-content target). But as of April 2026 the negotiation is reported to be stalled over precisely the sovereignty question: India is demanding access to the aircraft’s Interface Control Document — the blueprint for integrating its own weapons (Astra, Rudram, BrahMos-NG) without recurring French sign-off — and France is reportedly refusing, partly over fears that integrating the Indo-Russian BrahMos missile could leak sensitive avionics and electronic-warfare software to Russian entities. A foreign platform bought under an “Owned by India” doctrine, now snagged on the exact word “owned.”

The fuller picture

1. DAP 2026: what exactly changed

The Department of Defence released the draft Defence Acquisition Procedure 2026 on 10–11 February 2026, with feedback invited by 3 March 2026. The official objective is “to propel Jointness, Atmanirbharta & Integration, Force Modernisation, and Speed of acquisition with scaling of production.” Officials framed the philosophical shift as moving from “Made in India” to “Owned by India” — prioritising who owns the design, IP and source code over merely where equipment is assembled. The Observer Research Foundation (ORF) describes DAP 2026 as “the navigation chart for the middle leg of India’s journey to 2047.”

Category restructuring. DAP 2020’s five categories are reduced to four. The standalone “Buy (Indian)” route — which allowed procurement from an Indian vendor without indigenous design, provided content thresholds were met — is dropped, making “Buy (Indian-IDDM)” the primary domestic route. Critics (notably Amit Cowshish, former Financial Adviser for acquisition at the MoD) call the renaming of “Buy and Make (Indian)” to “Buy (Indian) and Manufacture in India” as cosmetic as the equivalent 2020 rebrand, since the defining characteristics are unchanged.

Indigenous content. Raised from 50% to at least 60% in Buy (Indian-IDDM), calculated against a defined “Calculated Base Contract Price.” At least 60% in the manufacturing portion for “Buy (Indian)” and “Buy (Global) & Manufacture in India.” Even “Buy (Global),” which previously had no obligation, now requires up to 30% indigenous content plus a parallel indigenous design-and-development pathway.

“Owned by India” / IP / source code. A system qualifies as indigenously designed only where the Indian entity owns the design documents, software source code, system architecture and critical technical data — not merely a transfer-of-technology licence. The aim is enduring control over upgrades, modifications and subsystem integration (and the freedom to export derivatives without the original manufacturer’s permission).

Foreign subsidiaries excluded. Wholly-owned Indian subsidiaries of foreign firms do not qualify as Indian vendors under the preferred Buy (Indian-IDDM) route — a deliberate shield for domestic players, but one that also closes the fastest route to the component ecosystem India lacks.

Speed mechanisms. Technology-readiness-level (TRL — a 1-to-9 scale of how battle-ready a technology is) categorisation for all routes except Buy (Global); single-vendor procurement allowed in Buy (Indian-IDDM) for mature (TRL 6–9) equipment; two-stage trials; a refined Fast Track Procedure with decisions delegated downward for emerging, short-cycle technologies; compensation for all vendors who clear trials; and timeline monitoring from the very first “Request for Information” stage. The draft explicitly notes that AI, drones, space and cyber technologies move in 18–36 month cycles — often shorter than a single procurement cycle.

iDEX/Make elevation and spiral development. iDEX (the Innovations for Defence Excellence startup programme), the Technology Development Fund and the “Make” route are elevated from peripheral status to mainstream acquisition pathways, with five years of assured orders and spiral (iterative) development. When a prototype passes trials, the sponsoring Service must place a meaningful Minimum Pilot Order Quantity. “Make-I” projects can receive up to 70% government funding (capped at ₹400 crore per Development Agency), 100% in select cases.

Low-Cost Capital Acquisition. A new pathway for fast-moving, low-cost items — an “exploit-first, scale-later” route to acquire equipment from Indian vendors for trial and evaluation before bulk induction, with a per-project threshold of around ₹75 crore and a ₹2,000 crore annual cap, powers delegated to the Services Procurement Board.

Exports. DAP 2026 sets a target of ₹50,000 crore in annual defence exports by 2030 and makes “exportability” a desirable qualitative requirement for new platforms.

Offsets vanish. The offset requirement — which had obliged foreign vendors to plough a share of contract value back into India — is conspicuous by its absence, replaced by the strengthened indigenous-content framework. (The history of why is a cautionary tale; see §6.)

Criticism of the draft. Cowshish argues the three-week comment window for an 800-page document is “more of a formality”; that several category renamings are cosmetic; that incentive/weightage systems (like the 2016 enhanced-performance-parameter scheme) have historically underperformed; and that the deeper structural problems — a multiplicity of agencies, the absence of a professional procurement cadre — remain unaddressed. ORF flags that the “indigenous design” test is “heavily technology- and documentation-centric, potentially excluding integrators who add value through hybrid designs incorporating foreign IP,” and that the whole Make-iDEX-LCCA ecosystem depends on the Services being willing to accept “limited-scale induction, iterative improvement, and occasional failure” — a cultural shift, not a procedural one.

2. Operation Sindoor: procurement and capability lessons

Operation Sindoor ran 7–10 May 2025, in response to the 22 April Pahalgam attack that killed 26 civilians. India struck nine terror-linked sites; Pakistan retaliated with drones and missiles against military installations across northern and western India.

Combat claims are disputed and should be flagged. India’s account claims neutralising hundreds of Pakistani drones and downing four-to-five Pakistani fighters plus a large airborne-early-warning aircraft. Pakistan claims it downed up to six Indian aircraft including three Rafales. India did not initially confirm losses; the Chief of Defence Staff later acknowledged unspecified losses but called Pakistan’s six-jet figure “absolutely incorrect,” stressing that India “rectified tactics” after 7 May and then struck airbases deep inside Pakistan. Independent assessments (a US Congressional commission, a Swiss think-tank, the French Air Force chief, and the strategist Bharat Karnad citing recovered wreckage) cluster around three Indian losses on 7 May — variously a Rafale, a MiG-29 and a Su-30. Treat all loss and kill figures as contested between the two sides.

The Emergency Procurement route delivered the tempo. The MoD’s sixth-phase Emergency Procurement concluded 13 contracts worth ₹1,981.90 crore against a ₹2,000 crore sanctioned outlay for the Army. Key items: integrated drone-detection-and-interdiction systems, low-level lightweight radars, very-short-range air-defence launchers and missiles, remotely piloted vehicles, loitering munitions (including vertical-take-off types), various drones, bulletproof jackets, ballistic helmets, quick-reaction fighting vehicles and rifle night sights. Separately, the Army ordered ~450 Nagastra-1R loitering munitions (>80% indigenous) from Solar Defence & Aerospace. Emergency-Procurement powers — first granted after the 2016 surgical strikes and reused after the 2019 Balakot strike — let the Services bypass the lengthy DAP, with authority to spend up to 15% of capital budgets this way. That the war-winning kit came through the bypass is the central diagnostic about the mainstream system.

Indigenous air defence performed. Akash (DRDO; ~96% indigenous) earned a “stellar” rating from the Director General of Military Operations; Akashteer (Bharat Electronics) networked radars for automated, decentralised engagement; and the integrated air-command system coordinated tri-service responses, knitting together the Indo-Israeli Barak-8 and the Russian S-400. The clearest doctrinal lesson is the “integration win” — sensors, datalinks, electronic warfare and interceptors operating as one loop, rather than any single platform.

Counter-drone systems. The D4 (DRDO/BEL — radar, radio-frequency, electro-optical and jamming, with lasers); SAKSHAM (an AI-based grid inducted October 2025); Bhargavastra (Solar Defence — micro-rockets and micro-missiles, developed without a formal tender); and DRDO laser directed-energy weapons (~2 km range).

The procurement surge. The Army moved to procure ~850 loitering “kamikaze” drones under fast-track, with reported plans for tens of thousands of loitering munitions across formations; each infantry battalion is to field a drone platoon. These figures come largely from media citing defence sources rather than official releases (see caveats).

The Defence Secretary’s diagnosis. Rajesh Kumar Singh called Sindoor a “reality check,” identifying capability gaps in “electronic warfare, counter-unmanned systems, better manufacturing ecosystem for military-grade drones, which can survive better in GPS-denied and other contested environments [and] various types of low-level radars.” He distinguished India’s large civilian drone ecosystem from its thin military-grade base (“three to four, five manufacturers”), pledged to use the government’s monopsony buying power to build it, and said procurement — even for complex items — should not exceed two years. He acknowledged that military-grade drones will need foreign technology tie-ups.

3. The domestic innovation ecosystem

iDEX. Launched in 2018. As of early 2025: 549 problem statements, 619 startups/MSMEs engaged, 430 contracts signed; grants run up to ₹1.5 crore (₹10 crore under iDEX Prime) and up to ₹25 crore for deep-tech under the ADITI scheme. The binding constraint has been the “valley of death” between prototype and order — exactly what DAP 2026’s assured-order provision is meant to bridge.

Drone ecosystem. Over 600 drone startups, more than $500 million raised collectively. Goods-and-services tax on drones was cut to a uniform 5% (from 18%/28%) effective 22 September 2025, with military-grade drones and high-performance batteries exempted entirely. A production-linked incentive scheme (₹120 crore) is being expanded toward component manufacturing.

The supply-chain dependency problem (the crux). Indigenisation by value reaches 70–80% on advanced Indian drones — but the strategically critical 20–30% (flight-controller chips, electro-optical sensors, high-performance motors, rare-earth magnets, batteries) is largely imported, mostly from China, which controls roughly two-thirds of global drone production. China’s grip on the upstream material is tighter still: it controls the overwhelming majority of global rare-earth processing and magnet manufacturing, and it restricts magnet exports for defence end-use. In 2023–24 the Army scrapped three contracts for 400 drones after discovering Chinese components in “Made in India” UAVs. This is the “paper-content on a fragile supply chain” risk in concrete form — and the reason a 60% indigenous-content floor can be satisfied while the war-critical fraction remains foreign.

Breaking the monopolies. Under DAP 2026’s direction, the MoD ended Bharat Dynamics Ltd’s tactical-missile production monopoly, designating private firms (Adani Defence, Bharat Forge, ICOMM, Solar Defence) as development-cum-production partners on DRDO missile programmes, alongside the public-sector incumbents in a hybrid model. This follows the breaking of Hindustan Aeronautics’ aerospace monopoly. The public sector still dominates production (~79% public / ~21% private), but the direction of travel is unmistakable.

4. The numbers

  • FY2026-27 defence budget: ₹7.85 lakh crore (highest ever; +15.19%; ~2% of GDP; the largest allocation of any ministry).
  • Capital: ₹2,19,306 crore (+21.84%), of which ₹1.39 lakh crore (75% of capital acquisition) is reserved for domestic industry.
  • The pension/revenue squeeze: defence pensions alone are ₹1.71 lakh crore for 34 lakh-plus pensioners (One Rank One Pension), and pay and allowances another large slice; capital is under 28% of the total budget. Modernisation competes for what’s left.
  • Contracts: ₹1.82 lakh crore of capital contracts in FY2025-26 (to December); Acceptances of Necessity worth ₹3.5–3.8 lakh crore since January 2025. Headline deals include 97 Tejas Mk1A fighters (₹62,370 crore) and 156 LCH Prachand helicopters.
  • Production: ₹1.78 lakh crore (record). Exports: ₹38,424 crore in FY2025-26 (+62.66%), to over 80 countries; the target is ₹50,000 crore by 2029-30.
  • Import ranking: Per SIPRI (the Stockholm International Peace Research Institute), India was the world’s 2nd-largest arms importer in 2021–25 at 8.2% of global imports — down 4% on 2016–20 as indigenous capability grew. Russia’s share of Indian imports fell to 40% (from 51% in 2016–20 and 70% in 2011–15), with France at 29% and Israel at 15%.

5. The 114-Rafale deal

India’s Defence Acquisition Council gave in-principle approval in February 2026 for 114 Multi-Role Fighter Aircraft, valued at ~₹3.25 lakh crore (~$40 billion) — India’s largest-ever arms acquisition if finalised. The structure: 18 flyaway aircraft from Dassault, 96 made in India, the F4 standard with a path to F5, a final assembly line in Nagpur, an engine maintenance facility, and a ≥60% indigenous-content target by end of production.

The IP tension is now an active impasse. As of April 2026, defence-trade media report the negotiation as stalled. India’s demand is reportedly not the entire source code but access to the Rafale’s Interface Control Document — the technical blueprint defining how weapons, sensors, electronic-warfare systems and mission computers talk to each other — without which Indian engineers cannot independently integrate indigenous weapons (the Astra air-to-air missile, the Rudram anti-radiation missile, BrahMos-NG) or update the SPECTRA electronic-warfare suite’s threat library as Chinese and Pakistani systems evolve. France is reportedly resisting, in part over fears that deep BrahMos integration could expose sensitive avionics to Russian entities tied to the BrahMos joint venture. The reported negotiating line — “No Interface Control Document, no deal” — is the single sharpest illustration of the whole argument: even a buyer spending $40 billion struggles to own what it buys. (Note: the “stalled/crisis” framing comes from defence-trade outlets, not official statements; the underlying technology-access terms are, by their nature, not public.)

6. Critical perspectives — both sides

The case FOR the sovereignty bet (why “Owned by India” is right even if slower). Foreign technology denial during conflict is real — the Defence Secretary’s own warning that “critical technology will not be shared, or shared at a very high rate” is the core argument, and the Rafale source-code refusal proves it. India has lived the Russia-spares-dependency trap that Ukraine’s war exposed, where imported, closed-source systems can be throttled by a supplier under geopolitical pressure. The 75% domestic earmark, the monopoly-breaking and design ownership build enduring capacity; SIPRI confirms imports falling as indigenous capability grows. And design/source-code ownership is what enables upgrades, mid-life refits and derivative exports without a foreign veto.

The case AGAINST (why the speed worries are real). Lt Gen Raj Shukla (Retd) argues India risks “preparing for the last and not the next war,” and calls for 40–50% of modernisation funds to go to “dronery, data readiness, sovereign AI [and] algorithmic warfare,” and for at least 30% of orders to go to startups — warning that “if at least 30% of the orders are not going to start-ups, our defence procurement pipelines are simply not futuristic and agile enough.” The strategist Bharat Karnad puts it more bluntly: in “this age of drones,” manned fighters are “an unaffordable luxury,” the Rafale is “entirely redundant to need” at ~₹2,850 crore per aircraft, and the same money should “flood the forces with thousands of attritable drones, counter-drone systems, and autonomous kill chains right now” while scaling up the indigenous Tejas through parallel private-sector production. Beyond the rhetoric, three structural points stand: the fact that Sindoor’s tempo came from Emergency Procurement is an admission the mainstream can’t keep pace; a 60% content floor can be met while war-critical chips, magnets and sensors stay Chinese (the scrapped-contracts episode and the 46-of-46 trial failure are the evidence); and China’s advantage is ecosystem scale — the Shenzhen model of dense component integration — which India cannot copy but must adapt through genuine civil-military fusion.

Offsets — a cautionary tale for content mandates. The Comptroller and Auditor General’s 2019 performance audit found that of 46 offset contracts worth ₹66,427 crore signed between 2007 and 2018, vendors had discharged only ₹11,396 crore, of which the MoD accepted just ₹5,457 crore — and not a single case of high-technology transfer to Indian industry had materialised. The flagship example: in the 36-aircraft Rafale offset, Dassault and the missile-maker MBDA proposed giving DRDO technology for the indigenous Kaveri engine, but “till date the vendor has not confirmed the transfer.” Offsets were progressively dropped for government-to-government, inter-governmental and single-vendor deals because they delivered no meaningful technology transfer; just one offset contract was reportedly signed between March 2021 and March 2025. DAP 2026 abandons offsets in favour of embedded content — but the offset graveyard is a warning that a mandate on paper is not capability in fact.

7. Historical and comparative context

The speed problem is structural, not new. India has rewritten this rulebook roughly every few years — Defence Procurement Procedures in 2002, 2005/06, 2008, 2011, 2013 and 2016, then DAP 2020 — and each iteration multiplied categories and procedural complexity without fixing the core delay drivers: the formulation of qualitative requirements, the trials regime, and contract negotiation. DAP 2026 dispenses with none of those three.

The lowest-bidder and over-specification problems persist. The “L1” system (award to the lowest technically-compliant bid) rewards cost-cutting over quality and ignores life-cycle cost; a former Army Chief publicly called to scrap it. Over-specified qualitative requirements routinely produce failed tenders — the classic case being a 2011 assault-rifle tender demanding a weapon that fired two different calibres of ammunition, which no vendor could meet, forcing an emergency import. These are exactly the frictions that push urgent needs into the Emergency Procurement channel.

Comparative models. Ukraine now iterates drones at wartime speed, redesigning around monthly attrition; the US “Replicator” initiative pursues attritable mass; China’s civil-military fusion and Shenzhen component base give it scale India lacks upstream. The common thread across all three is that tempo and component sovereignty, not platform prestige, decide drone-age conflicts.

How to fix it

  1. Carve out the software-speed categories. Ringfence drones, counter-drone systems, electronic warfare and loitering munitions from the 60% indigenous-content floor and the documentation-heavy “indigenous design” test built for big platforms. For these categories, prioritise fielding speed and iteration cadence over paper-content — accept iterative improvement and occasional failure. Marker that would change this: once domestic component manufacturing (chips, magnets, motors, batteries) reaches genuine scale, tighten the content floor back up.

  2. Convert Emergency Procurement’s tempo into the mainstream default for fast-cycle items. The Emergency route worked because it ran steps concurrently rather than sequentially. Institutionalise that concurrency — not its emergency-only status — for sub-₹75-crore Low-Cost Capital Acquisitions and iDEX-origin items, and hold the Defence Secretary’s stated two-year ceiling as a hard, published service-level commitment tracked from the first Request for Information.

  3. Treat component sovereignty — not assembly — as the real indigenisation metric. Fund and accelerate domestic production of flight-controller silicon, rare-earth magnets, electro-optical sensors and drone-grade batteries (the IIT-Kanpur hub is a start, not a finish). Threshold: until India can produce these at consistent tolerance and cost at volume, a high-paper-content drone is renting capability, not owning it. The 46-of-46 trial failure is the clearest current signal that the binding gap is upstream, not in airframes.

  4. Hit the 30% startup-order benchmark, and publish it. Use the assured-order and Minimum Pilot Order Quantity provisions to direct a measurable share of fast-cycle orders to startups and small firms, and report that share annually.

  5. Be honest about the Rafale exception. A ₹3.25-lakh-crore foreign platform without integration rights is defensible as a squadron-shortfall stopgap — but it should be explicitly bounded, with the Interface Control Document fight used to extract maximum integration sovereignty, and not allowed to crowd out the drone, electronic-warfare and counter-drone spend that Sindoor showed is decisive.

Caveats

  • DAP 2026 is a draft, not notified law (comments closed 3 March 2026; the aerospace chapter is still pending). The provisions described here are proposals; provision-level detail is corroborated via MoD/PIB releases and analyses by ORF, ThePrint and law firms, rather than direct quotation of the ~800-page text.
  • Operation Sindoor combat claims are disputed. Indian aircraft losses were claimed by Pakistan; India acknowledged unspecified losses while rejecting Pakistan’s specific numbers. Independent assessments cluster around three Indian losses; figures on both sides should be treated as contested.
  • The 46-manufacturer trial finding originates in an India Today investigation and is corroborated across several defence outlets, but has not been confirmed in an official MoD release. Note the important nuance that a harder October re-test was passed by a handful of firms — the story is “the Army raised the bar to combat-realistic electronic warfare and most, not all, firms initially failed,” not “Indian drones are hopeless.”
  • The Rafale “stalled/crisis” framing comes from defence-trade media, not official statements; technology-access terms are not public.
  • Much post-Sindoor procurement detail (the 850 drones, tens of thousands of loitering munitions) is reported by media citing anonymous defence sources, not always confirmed officially.
  • Several figures vary by source and reporting date (₹1.82 vs ₹2.10 lakh crore of contracts; $36 vs $40 billion for Rafale; SIPRI’s 8.2% for 2021–25 vs 8.3% for the prior 2020–24 fact sheet). The export figure used here is the FY2025-26 record of ₹38,424 crore, which supersedes the earlier ₹23,622 crore (FY2024-25) number.

Sources

All URLs accessed 21 June 2026. Listed in priority order — primary/official, then wire services and papers of record, then domestic outlets across the spectrum, then think-tank and specialist analysis, then data, with encyclopedias treated as a map only. Where the article relies on a single outlet for a contested characterisation (notably the Rafale “stall” and the drone-trial failure), that is flagged in the Caveats above.

Primary / official (documents, filings, official statements)

Wire services / papers of record

Domestic outlets (across the spectrum)

Think-tank / specialist analysis (framework)

Data / reference guides (aggregators — used with caution, cross-checked)

Map only (navigated, not cited as authority)