India Defence Tech
Deep Dive
Defence stopped being a government business and became a venture category. India is the only major market that closed its own market to imports - and then built funded rails to route startups into the demand it created.
Authors
Pooja Akkem
Prashant Rathee
Gitika Velichala
Nishaanth Rangan
Setup
Foreword: Why This, Why Now
In 2025 the world spent $2.89 trillion on defence - the eleventh consecutive year of increases and the highest total ever recorded. But the headline flatters a more interesting shift underneath it. For the first time in years, the largest spender shrank: US military spending fell 7.5%. It didn't slow the world down. Strip out the United States and everyone else's spending rose 9.2%, with Europe up 14% and Asia and Oceania up 8.1%.
The money is decentralising. So is the innovation. For most of the last thirty years defence was the wrong shape for venture capital - slow procurement, a handful of dominant primes, and a single government buyer that took a decade to say yes. What has emerged instead is plural and structural: sustained, multi-decade demand across dozens of countries, each with its own sovereignty drivers, and a supply base where venture-backed companies are increasingly the ones building the decisive systems.
India sits at the sharpest point of that shift. It commits 8.2% of its government budget to defence - almost exactly the United States' 8.3% - while being a fraction as rich. It has legally foreclosed foreign supply across roughly 5,500 defence items. And in May 2025 it fought a war that was won on indigenous counter-drone and air-defence systems, converting policy intent into a hard, funded demand signal. This report argues that the combination is a once-a-decade opening, and maps where it is investable.
Setup
Who This Is For
This is written for two readers. Founders building deep technology with a defence or dual-use surface, who need to know whether the Indian demand signal is real and how to reach it. And investors - LPs and co-investors - underwriting the India sovereign-tech thesis, who need to know how a government anchor buyer changes underwriting, timelines and exit maths.
Defining Indian Defence Tech
We treat as in-scope any business with the Ministry of Defence or the armed forces as a core growth pillar - whether it sells only to government, sells dual-use into both defence and commercial markets, or starts in defence and expands outward. The definition is deliberately broad, because the most investable companies in this category rarely stay on one side of the line.
Setup
What You'll Walk Away Knowing
Claim 01
Defence is now venture-scale. Capital is moving from public primes into private hands structurally, not cyclically.
Claim 02
India hasn't merely favoured domestic defence startups - it has legally foreclosed the foreign alternative. No other major market combines import bans this broad with startup-specific demand rails. This is forced localisation, not a size-based preference like the US Small Business Act.
Claim 03
A government anchor buyer rewrites the risk maths. It removes the first-customer problem and compresses the path to scale and exit.
Claim 04
Defence deep-tech is dual-use by default, so winners become critical national infrastructure - widening the buyer set and the exit surface.
Part I
The Structural Moment
The argument in this part runs top-down. First the global picture: how much is being spent, by whom, and how fast it is growing. Then the United States, as the template for how a defence economy hands work to private hands. Then India, read through the same lens.
Global: The Money Got Serious, Then It Moved
World military expenditure reached $2,887 billion in 2025, up 2.9% in real terms and 41% across the decade. The top five spenders - the United States, China, Russia, Germany and India - accounted for $1,686 billion, or 58% of the global total.
$954B
US, 2025
−7.5%
$336B
China
+7.4%
$190B
Russia
+5.9%
$92.1B
India
+8.9% · 5th globally
The distribution matters more than the total. The US share of world spending has fallen steadily since 2020, dropping 4.3 percentage points in 2025 alone; it now outspends China by 2.8 times, down from 3.2 a year earlier. Growth has relocated to Europe, up 14%, and to Asia and Oceania, up 8.1%. A defence-technology market that was effectively one buyer deep is becoming many.
Two Ways to Read National Commitment
The first is share of GDP - economic weight. By this measure India is moderate: 2.3% of GDP, level with Germany, just under the United Kingdom, comfortably past the NATO 2% line but nowhere near war economies like Russia (7.5%) or Saudi Arabia (6.5%). The inference is headroom - India is not close to a wartime ceiling, so spending can keep rising for years without straining the economy.
The second is share of the government budget - political priority. Here India looks entirely different. At 8.2% of general government spending it sits almost exactly level with the United States at 8.3%, and well above the United Kingdom (5.5%), China (5.1%), Germany (4.5%), Japan (3.8%) and France (3.5%).
India commits to defence like a superpower and spends like a developing economy - moderate burden, maximal priority.
Read together, the two measures say something durable. Within a constrained budget and against enormous competing claims, India still places defence near the front of the queue. That is a revealed preference, not an accounting accident - and it is the reason this demand should be underwritten as structural rather than discretionary.
The United States: What a Mature Defence Economy Hands to Private Hands
The US spent about $951 billion on defence in FY2026 - a figure larger than the GDP of all but around twenty countries. The Pentagon does not merely buy: it shapes demand, underwrites the supply chain, funds innovation and directs the flow of capital through the entire defence economy. Understanding how it does that is the template against which India's system should be read.
The demand itself is changing shape. The One Big Beautiful Bill reconciliation act, passed in July 2025, pushed defence appropriations out through FY2029 - removing the annual time constraint and enabling multi-year research, development and supply-chain programmes. The Trump administration split defence into discretionary and mandatory funding under a secondary Department of War. The FY2026 request then set the emphasis clearly: research, development, test and evaluation rose 27% and procurement rose 18%, against roughly 5% for operations and maintenance. The message is unambiguous - the priority is enabling innovation and scaling new technologies, not near-term weapons production.
Roughly 40–45% of all federal contract spending flows to the top 25 US defence and aerospace companies, sustaining a non-governmental supplier economy of around $300 billion a year. The top five - Lockheed Martin, RTX, Northrop Grumman, Boeing Defense and General Dynamics - booked about $202 billion from defence last year between them.
The disruptors are a small group growing very fast. Anduril - the only venture-backed company approaching the top 25 - along with Palantir and SpaceX, are compounding five to twenty times faster than the legacy primes, building new supply chains from scratch in which subcontracting suits smaller companies.
What makes this reachable for a startup is that the US government has built four distinct lanes into its money, and each carries a different signal:
The de-risking signal is what matters. When the government tells you what to build, builds it alongside you, and then buys it, the residual risk for an investor is execution and timely scaling rather than technology or market discovery. That is a fundamentally different underwriting problem - and it is the mechanism this report argues India has now replicated, by a different route.
That continuous involvement also catalyses exits. VC exits in US defence surged to $54.4 billion in 2025, up from $18.2 billion in 2024, most of it through acquisition - and the direction is telling: government-backed incumbents are increasingly buying venture-backed companies rather than building internally. Risk has moved from technology to supply chain and execution; innovation is decentralised, while production becomes the controlled channel.
India: Small by Dollars, but the Curve Is Bending
Read India through the same venture lens and the numbers look tiny - until you see the structure underneath them. On a consistent equity-funding measure, global VC into defence startups hit $17.9 billion in 2025, about 2.5 times the $7.3 billion of 2024, making it one of the fastest-growing categories in all of venture capital in a year when funding outside AI was largely flat. Roughly ten new defence-tech unicorns were minted, taking the cohort's combined valuation to around $495 billion.
The dollars are highly concentrated. The United States took about $14.2 billion of the 2025 total - roughly 79% - Europe $2.48 billion (~14%), the rest of the world about $1 billion (~6%), and India roughly $0.25 billion, about 1.4%. A single US deal makes the point: Anduril's $2.5 billion Series G, at a $30.5 billion valuation, was larger than all of Europe's defence VC for the year and roughly ten times everything India raised.
$17.9B
Global defence VC, 2025
2.5× 2024
79%
Captured by the US
$247M
India - record year
~30
India rounds
India's $247 million was an all-time high, roughly double the prior year, but heavily concentrated: Raphe mPhibr's $100 million Series B (drone swarms, led by General Catalyst) was about 40% of the entire year on its own; NewSpace Research (~$33 million) and Sagar Defence (~$25 million) took the top three to about 64%, with some 27 further rounds sharing the remaining ~$89 million across ISR, space, counter-drone and semiconductors. The funding skews strongly to dual-use, non-combat systems rather than weapons.
The trigger was a war. Raphe mPhibr's drones were deployed in Operation Sindoor, and its round closed only weeks afterward; in the conflict's wake India tripled drone-related spending to around $470 million. It is the same sequence Ukraine set off for US funding - battlefield validation first, capital second. (Trackers diverge on the India total; we use Tracxn's $247M as the most-cited record figure that reconciles with the named rounds.)
Everywhere else, capital bets that demand will show up. In India, demand is the one thing already certain.
That is why the small funding number understates the opportunity. In the US and Europe, venture takes the first risk, funding a startup that must then win a contract in an open, crowded market. In India the sequence is inverted: the government de-risks demand before private capital arrives - through iDEX and ADITI, and through a buy-local rule that directed roughly 81% of procurement value to Indian industry in FY25. The $247 million of private money sits on top of demand that is already guaranteed. That floor is what the rest of this report is about.
Part II
Why India Is Different: Forced Localisation, Not Preference
Every major defence market wants more procurement to reach domestic and smaller firms. The methods, though, are variations on a single theme - help startups reach an open market. India pulls a different lever entirely.
United States
Size-based set-aside
The Small Business Act reserves ~23% of contract dollars for small firms; the Rule of Two forces set-asides. SBIR/STTR and OTAs cross the valley of death. But every lever keys on size, not origin - a foreign-owned small firm can still bid.
→ a reserved slice
France / EU
Access and capital
Europe's best startup front door: AID, a single desk to pitch the military, plus RAPID grants, Definvest, ~€5bn mobilised in 2025. But France buys inside EU procurement rules and joint programmes. It can streamline access; it cannot ban imports.
→ a front door
India
Import ban, plus rails
The Positive Indigenisation Lists prohibit foreign supply across ~5,500 items. iDEX and ADITI then fund startups into that mandated demand. Not a share of an open contract - the removal of the foreign bid, plus a funded route to fill it.
→ a guaranteed floor
How India Closed the Market
The Positive Indigenisation Lists, introduced under the Defence Acquisition Procedure 2020, do not reserve a share. They remove the alternative. Each item carries a cutoff date; once it passes, that item can only be bought from Indian industry, with staggered import-ban dates extending to 2028.
The mechanism runs on two tracks. The Department of Military Affairs has notified 509 items for the armed forces - complex systems, sensors, weapons and ammunition. The Department of Defence Production has notified close to 5,000 more for the defence public sector undertakings, largely sub-systems, spares and components. Nearly 5,500 items in total, foreclosed to foreign supply - of which about 3,000 had been indigenised by February 2025.
This is the Buy Indian–IDDM logic of DAP taken to its limit. The scale of intent is real: the SRIJAN portal lists over 34,000 items for indigenisation, with more than 10,000 already indigenised. And the budget follows it - ₹1.39 lakh crore, 75% of this year's capital acquisition budget, is earmarked for domestic industry, private players included.
₹7.85L cr
FY26–27 budget · +15% · top ministry
~5,500
Items foreclosed to imports
₹1.39L cr
75% of capital acq., domestic
81%
Of FY25 contract value, domestic
How Startups Actually Reach That Demand
A ban alone would not help a startup. A state-owned giant could absorb every order. So India built the rails, and structured them as a ladder:
One precision matters, and the report holds to it throughout: the lists mandate buying Indian, not buying from startups. A defence PSU can fill the order. What iDEX and ADITI add is a funded, structured path for a startup to compete for demand already guaranteed to stay onshore. The combination is the claim - a legally protected domestic market, plus rails built specifically to route new entrants into it.
The US and France help startups reach the market. India removed the foreign bid - and funds startups to fill it.
Case Study · May 2025
Operation Sindoor: When Self-Reliance Met the Battlefield
In April 2025 a terror attack on tourists in Pahalgam killed 26 people. India's response, Operation Sindoor, became the clearest real-world test yet of its self-reliance push.
On the night of 7–8 May, Indian forces struck terror infrastructure inside Pakistan without crossing the Line of Control or the international boundary. It was fought as a non-contact war - a tri-service deployment of drones and missiles for precision strikes at depth, rather than massed formations at the border. The offensive drew on both imported and indigenous stand-off weapons: SCALP and HAMMER munitions alongside BrahMos, the Akash air-defence system and SkyStriker loitering munitions.
The decisive proof came on defence. Across the nights that followed, Pakistan attempted strikes on some fifteen military installations across northern and western India. These were neutralised by India's Integrated Counter-UAS Grid and air-defence systems - largely home-built, knitted together by the AI-driven Akashteer and IACCS networks. The Director General of Air Operations stated that waves of Pakistani drones and unmanned combat aircraft were thwarted by indigenously developed soft- and hard-kill counter-UAS systems, with Akash performing notably. The Chief of Defence Staff later confirmed that none of Pakistan's drones inflicted damage on Indian military or civil infrastructure.
Two lessons matter. First, the establishment's own reading was unambiguous: the CDS said the operation underscored the critical importance of indigenous counter-UAS capability, and that dependence on foreign technology weakens preparedness. The conviction translated into money - military aircraft capital outlays were revised 50% above budget during the year.
Second, the war validated exactly the kind of technology a venture portfolio can build. What decided the outcome was not a $100-million platform but drones, loitering munitions, counter-UAS, sensors and the software that fused them. Sindoor turned India's drone and autonomy ecosystem from promising into proven - and converted political intent into a funded demand signal for precisely the segments mapped in Part III.
Part III
The Indian Landscape Market Map
Indian defence tech is not one market. It is a set of technical layers that come together inside a larger system: platforms, electronics, sensors, autonomy, software, communications and protection. The categories overlap - a UAV company depends on flight-control electronics, payloads, autonomy software and secure comms - but separating the layers clarifies where Indian companies are actually building capability.
The map is best read on two axes. Enablers - AI and autonomy, semiconductors and electronics, sensors and ISR, cyber and secure communications - run across domains: air and aerospace, land, sea, space. Aerospace is the domain where India's base is deepest; the enablers are where the venture-scale gaps sit.
The US map is software-heavy because it already owns the hardware base. India's inverts: the gap is hardware sovereignty, plus the same software layers everyone needs.
Aerospace & Space, The Domain Anchor
The most visible and most mature part of the sector: UAVs, loitering systems, launch vehicles, satellite platforms, payloads, propulsion and unmanned maritime platforms. ideaForge and Raphe mPhibr represent the established UAV layer; Apollyon Dynamics is part of the newer loitering-UAV cohort. The more important development is the growing set of companies building difficult subsystems rather than only end platforms - Nautical Wings (axial-flux motors, composite structures), Nabhdrishti and Dheya (micro gas turbines and turbojets), Astrome and XDLINX (satellite comms and payloads). The airframe is only one layer; propulsion, avionics, payloads, navigation, structures and power determine how much can be built locally.
Hardware Sovereignty, Semiconductors, Electronics & Materials
Every category eventually comes back to electronics, and this is the deepest gap and the hardest moat. India already has depth in RF and microwave systems, embedded electronics, radar subsystems and defence-grade integration through Data Patterns, Astra Microwave, Centum, Avantel and Saankhya Labs. The newer cohort builds the layers autonomous systems depend on: Yaanendriya (IMUs, AHRS, GNSS, flight-control hardware), ChipUAV (flight controllers), AGNIT (GaN RF), Netrasemi (edge-AI silicon), Sanlayan (radar, EW, avionics, power), and RF Nanocomposites (radar-absorbent materials, EMI shielding).
The gaps are real and openly acknowledged
Propulsion is the single most important technology India cannot yet produce at scale: completed Tejas airframes have sat idle awaiting GE F404 engines, and a follow-on $1 billion order for 113 more engines was placed in November 2025. On semiconductors, when India ordered Rafale jets the offset request for GaN chip-technology transfer was denied; DRDO's indigenisation with GAETEC has produced satellite-grade but not yet fighter-grade GaN, while India imported roughly $30.3 billion of semiconductors in FY25-26, up 23.3% year on year. Advanced materials - titanium, defence-grade carbon fibre, special alloys - remain largely imported, though a titanium and superalloy plant was established in Lucknow in October 2025. These are the gaps the Positive Indigenisation Lists are now forcing open on a timeline.
The Enabler Layer, Autonomy, Sensors & ISR, Cyber
Talent-driven rather than fab-driven, low in capital intensity, dual-use by default, and pulled by demand the battlefield has made impossible to ignore - this is the most fundable corner of Indian defence tech, because India's base is strongest in platforms and weakest in the software that makes them lethal.
Autonomy
The airframe was never the hard part - the stack that navigates, targets and operates when GPS is jammed and the link is cut is largely imported or absent. Autonomy is software, not fabrication, drawn from India's deep commercial robotics and self-driving talent. Shield AI has brought its Hivemind autonomy to the Indian Army; NewSpace Research (swarms), Sagar Defence (maritime and underwater) and Raphe mPhibr build into the same demand. The white space is sovereign, GPS-denied autonomy built for Indian conditions.
Sensors & ISR
Modern conflict is a see-first contest - putting up the sensors, and turning what they capture into a target in seconds. The hard part is the AI, and the skills transfer from India's fast-growing space and computer-vision ecosystem; the government is actively pulling private capacity in, with the ₹26,968-crore Space-Based Surveillance programme fielding 52 satellites, 31 from private firms. GalaxEye and Pixxel build SAR and hyperspectral satellites; Tonbo Imaging and Optimized Electrotech build the electro-optics; SatSure turns imagery into intelligence. The white space is the fusion layer.
Cyber & secure comms
The network is the first target in any modern war - jam it, blind it or break in, and the platforms above are useless. India runs one of the world's largest commercial cybersecurity industries, and the doctrine to pull it in is in place: a 2024 Joint Doctrine for Cyberspace Operations and an operational Defence Cyber Agency. QNu Labs (quantum-safe crypto), FleetRF (anti-jam drone comms), Vehere and Indrajaal (counter-UAS sensing and response) are early. This is the least-built layer of the three - which is exactly what makes it the largest opportunity.
Across all three, the same pattern holds: India already has the civilian version of the capability at world scale, so the defence version starts from a running base rather than a standing start, and real battlefield demand is pulling it forward. Because each is dual-use by default, a winner does not just serve the military - it becomes infrastructure the country cannot do without.
Part IV
The De-Risking Playbook
Parts I to III establish that the demand exists, that it is legally protected, and where it concentrates. This part addresses the question an investment committee actually asks: what does that do to the risk maths?
The Anchor Buyer Removes the Hardest Problem
In deep tech, the first major customer is usually the binding constraint - not the technology. A government that publishes its problem statements, funds the prototype, and commits to buy the output collapses three distinct risks - market discovery, first revenue, reference customer - into one: execution. That is the same transformation the US achieved through OTAs and SBIR, arrived at in India by the opposite route: mandated demand first, rails second.
The India Government-Sales Playbook
The founder-facing mechanics: iDEX challenge and Prime tiers, ADITI for frontier technology, TDF via DRDO, and the DAP acquisition categories - Buy Indian–IDDM at the top of the preference order - that determine whether a contract can legally be awarded abroad at all. Under DAP 2026 the bar rises further: equipment must be indigenously designed, developed and manufactured, and wholly-owned Indian subsidiaries of foreign firms no longer qualify.
Exits, Global PE, and Follow-On Capital
The direction of travel in the US is instructive and, we argue, leading: government-backed incumbents are increasingly acquiring venture-backed companies rather than building internally - the $54.4 billion of 2025 exits came mostly through acquisition. Risk has moved from technology to supply chain and execution, so innovation is decentralised while production remains the controlled channel. The open question for India is whether the same acquisition behaviour emerges from its primes and PSUs - HAL, BEL, Mazagon Dock, GRSE - and on what timeline. India's follow-on market is still thin: with the top three rounds making up ~64% of 2025, the Series B-and-beyond layer is where the funding gap, and the opportunity, now sits.
Dual-Use, and Why Winners Become Infrastructure
A company whose systems sit inside national-security architecture becomes strategically difficult to sell abroad and strategically valuable to hold. That constrains one exit route and widens several others - domestic primes, allied governments, and the commercial markets the same technology addresses. Claim 4 rests on this: the dual-use surface is what turns a defence contract into a platform.
Close
Conclusion and Implications
The through-line is supply-chain sovereignty. Global defence spending is decentralising, and so is the innovation that serves it - the decisive systems in both Ukraine and Sindoor were not generational platforms but attritable, software-defined, largely startup-built technology. That is the shift that made defence venture-fundable in the first place.
India's position within that shift is structurally distinct. It has the demand (₹7.85 lakh crore, the largest allocation of any ministry), the political durability (8.2% of government spending, level with the United States), the legal floor (~5,500 items foreclosed to imports), the rails (iDEX, ADITI, TDF), and now the battlefield proof. Few markets have any one of these. No other major market has assembled all five.
When a protected, prioritised demand floor meets venture capital and ground-up engineering, the result is not incremental. It is a new cohort of companies - and the window to underwrite them at entry prices is open now, not indefinitely.
For Founders
Build where the list already bans the import - the demand question is answered before you start. Use iDEX to fund the prototype and ADITI where the technology is frontier. Design for dual-use from day one: it is what converts a defence contract into a company. And treat the PSU as a competitor for the same mandated demand, not a distant incumbent.
For Investors
The anchor buyer converts market risk into execution risk, which is the risk venture is actually good at pricing. Underwrite the procurement pathway as carefully as the technology. Expect longer sales cycles but higher revenue durability, and price the dual-use surface - it is where the exit optionality lives.
| Rank | Country | US$ bn | Change | % GDP | % Govt |
|---|---|---|---|---|---|
| 1 | United States | 954 | −7.5% | 3.1 | 8.25 |
| 2 | China | 336 | +7.4% | 1.7 | 5.12 |
| 3 | Russia | 190 | +5.9% | 7.5 | 19.83 |
| 4 | Germany | 114 | +24% | 2.3 | 4.50 |
| 5 | India | 92.1 | +8.9% | 2.3 | 8.18 |
| 6 | United Kingdom | 89.0 | −2.0% | 2.4 | 5.45 |
| 7 | Ukraine | 84.1 | +20% | 40 | 63 |
| 8 | Saudi Arabia | 83.2 | +1.4% | 6.5 | 23.53 |
| 9 | France | 68.0 | +1.5% | 2.0 | 3.53 |
| 10 | Japan | 62.2 | +9.7% | 1.4 | 3.81 |
World total $2,887bn · top five $1,686bn (58%) · US 33% of world total. Source: SIPRI.
| Instrument | Scope | Scale |
|---|---|---|
| PIL - Dept. of Military Affairs | Armed forces: systems, sensors, weapons | 509 items |
| PIL - Dept. of Defence Production | DPSUs: sub-systems, spares, components | ~5,000 items |
| Capital acquisition earmark | Domestic industry incl. private | ₹1.39 lakh cr (75%) |
| iDEX | Startup / MSME challenge and Prime tiers | 430+ contracts · ₹2,400 cr+ |
| ADITI | ~30 frontier technologies | ₹750 cr · to ₹25 cr/project |
| TDF (DRDO) | Parallel development channel | - |
| Region | 2025 | Share | Note |
|---|---|---|---|
| United States | $14.2B | 79% | Anduril $2.5B Series G alone > all Europe |
| Europe | $2.48B | 14% | Breadth: funded startups +67% |
| Rest of world | ~$1B | 6% | Israel, South Korea, Australia |
| India | $0.25B | 1.4% | Record year; ~30 rounds |
Global total $17.9B (2.5× 2024), CB Insights equity series. India detail from Tracxn.
Global expenditure figures are from the SIPRI Military Expenditure Database (1949–2025) and its annual fact sheets, in current US dollars by calendar year, compiled from open sources. Defence-VC totals use a consistent equity-funding series ($7.3B 2024 → $17.9B 2025); trackers differ widely on defence boundaries, and India's $247M uses Tracxn as the most-cited figure that reconciles with named rounds.
A note on two denominators. India's defence spending is 8.2% of general government expenditure (Centre plus states) on SIPRI's basis, and approximately 14.7% of the Union Budget alone on the Ministry of Finance's basis. Both are correct; they measure against different bases. This report uses the SIPRI figure for cross-country comparison and labels the Union Budget figure explicitly wherever it appears.
On addressable spend. Salaries and pensions account for roughly half of India's defence budget, keeping the capital or modernisation share below 30%. The share genuinely addressable by new entrants is therefore materially smaller than headline allocations imply; any sizing in Part III should be built from the capital line, not the total.