Indian deep-tech in 2026 looks structurally similar to American semiconductor venture in 1985. The cost of capital is low. The talent pipeline is strong and reverse-flowing from global majors. The state has organised behind the sector. The strategic acquirers are domestic and increasingly visible buyers. Public-market exit windows are open. The founder mindset has shifted from defending a thesis to executing on it. We have backed ideaForge (NSE IPO 2023), Agnikul Cosmos, Bellatrix Aerospace, Detect Technologies, Astranova, PierSight, Dashamlabs, and others. The wave that is now arriving will produce more category leaders than the previous fifteen years combined.
What the team has actually catalyzed.
From IIMA Ventures, the Bharat Innovation Fund, Infuse Ventures, ADB Ventures, and the Bharat Inclusion Seed Fund - across space, defence, semiconductors, robotics, and materials.
ideaForge
India's largest drone manufacturer by market share. Deployed across Army, Navy, Air Force, and DRDO. NSE-listed since 2023 (₹567 crore IPO). IIMA Ventures portfolio.
Agnikul Cosmos
World's first 3D-printed single-piece rocket engine. IIT Madras spin-out. Series C at ~USD 500M valuation.
Bellatrix Aerospace
Electric propulsion and green propellants. IIMA Ventures and Bharat Innovation Fund portfolio.
Detect Technologies
Industrial AI for refineries, pipelines, and fired heaters. IIT Madras alumni; Accel and Prosus post-BIF.
PierSight
AI-powered satellite-image analytics. ISRO-SAC and IIMA ecosystem.
Dashamlabs
India's first silica-aerogel flexible insulation manufacturer. IIMA-V and Speciale Invest seed.
GalaxEye
World's first operational OptoSAR satellite. Mission Drishti launched May 2026 from IIT Madras lineage.
CynLr
Visual-intelligence-for-industrial-robotics platform. IISc-spin-out; USD 10M Series A November 2024.
ePlane Company
India's first DGCA-approved electric VTOL platform. Series B at ~USD 46M. IIT Madras Research Park.
Four numbers that set the frame
Defence-tech in India raised ~USD 192M in 2025 alone, 61× the 2016 number. Raphe mPhibr's USD 100M Series B is the largest Indian defence-tech round to date. iDEX has 600+ start-ups onboarded, 40+ prototypes cleared, and 31 procurement contracts totalling more than ₹1,560 crore awarded. ADITI 2.0 stacks grants of up to ₹25 crore per start-up.
The India Semiconductor Mission has approved 12 projects across 6 states, with cumulative approved outlay of ~₹1.64 lakh crore. The Tata-PSMC Dholera fab targets first silicon in late 2026. Micron's Sanand ATMP achieved first packaging in February 2026. The Union Budget 2026-27 allocated ₹8,000 crore in a single year to the semiconductor mission.
The Research, Development and Innovation Scheme - ₹1 lakh crore outlay - was approved in 2025 and is being operationalised through the Anusandhan National Research Foundation, with the ANRF itself having a ₹50,000 crore corpus. This is the largest single commitment to translational research the Indian state has ever made.
The defence-export target is ₹50,000 crore by 2029, up from ~₹21,000 crore in FY24. The Positive Indigenisation List of 411 items has created structural demand for Indian-designed and Indian-manufactured defence systems.
How the global story is changing
Capital has moved into deep-tech globally at unprecedented scale. Anduril at USD 61B, Shield AI at USD 12.7B, Saronic at USD 9.25B, Helsing at USD 18B. SpaceX is reported toward a USD 1.75–2T valuation. Rocket Lab is at USD 49B market cap. Figure AI closed Series C at USD 39B post-money. These are demonstrations that the deep-tech category at scale produces the largest venture outcomes of our generation.
The geopolitical realignment has made allied-supply-chain investment a strategic priority. Quad supply-chain alignment is real and operational. Indian listed defence stocks have re-rated 10–30× over five years. Pixxel's US NRO contract in May 2026 confirms that Western strategic capital is actively backing Indian deep-tech.
The open-source disruption in foundation AI has reduced the capex required to enter several deep-tech categories. Robot foundation models, satellite imagery analytics, generative materials models - the historical capex moats are shrinking, and Indian founders with serious science can now play at the global frontier.
The five sub-themes we invest in
Space. Agnikul, GalaxEye, Bellatrix, KaleidEO, PierSight, Orbitt. We invest in small-sat launch components, specialised payloads, in-space servicing, downstream EO analytics, and beyond-LEO infrastructure. 8–10 seed positions expected in this sub-theme.
Defence-tech. Indigenous UAV platforms with iDEX-led offtake, autonomous swarms, loitering munitions, ISR optics, unmanned ground and underwater vehicles, defence-grade materials, and dual-use AI. The access to Defence-Tech corridor at IIT Kanpur is our advantage.
Semiconductors. RISC-V CPU and microcontroller IP, edge AI accelerators, analog/mixed-signal IP, MEMS, GaN, SiC, silicon photonics, ATMP packaging, and EDA tools. We invest at the design and IP layer - not at front-end fab scale-up.
Advanced manufacturing and robotics. Vision-language-action robot foundation models for Indian factories, industrial AI for predictive maintenance, robotic test automation, additive manufacturing, eVTOL, and mobile robotics.
Materials and industrial chemistry. Battery materials beyond lithium-ion (sodium-ion, solid-state, flow chemistry), specialty alloys, single-walled carbon nanotubes, silica aerogel, bio-based materials, and specialty polymers from the Venture Center Pune pipeline.
What does not work
Pure-play cell scale-ups without anchor offtake - the Log9 cautionary tale is binding.
Pure semiconductor manufacturing scale-up rounds - Indian fab construction is being underwritten by central government commitments, large industrial groups (Tata, Vedanta, CG Power, Micron), and strategic foreign capital (Foxconn, PSMC). Private venture capital is not the binding constraint.
Single-platform space launch companies attempting to compete head-on with SpaceX on cost per kilogram to orbit. Skyroot and Agnikul occupy specific niches; we will not back a third Indian small-launch start-up.
The contrarian view we hold
The consensus that defence-tech is over-funded at the late stage is true for the US. It is not true for India. Indian defence-tech raised USD 192M in 2025 against a structural demand environment that supports an order of magnitude more capital. The Indian late-stage will be 1/5 to 1/10 of US late-stage in dollar terms but with comparable revenue and growth profiles - a structurally attractive entry point for institutional capital.
The consensus that Indian semiconductors will not produce venture-scale outcomes is wrong, narrowly. The design and IP layer (Mindgrove, InCore, Shakti-derived) has a structural opportunity to build IP that NVIDIA, Intel, and Qualcomm cannot match on price. The specialty-chip layer (GaN, SiC, photonics, MEMS) is structurally undersupplied globally.
Why full-stack wins in Indian hard-tech
Indian industrial buyers - listed OEMs, defence customers, infrastructure operators, public-sector enterprises - buy capability and uptime, not parts. A pure-product hard-tech company sells one cycle of equipment and loses the next ten years of recurring value to whoever runs the integration, the data, the spare-parts cycle, and the upgrades. The most successful deep-tech companies in our portfolio have built operating layers on top of their technology - and that operating layer is where the venture-scale value capture happens.
Concrete examples. ideaForge sells UAVs paired with surveillance-data services to defence, mining, and public-safety customers. Pixxel and Bellatrix do not just build satellites and launchers - they sell imagery and orbital services. Detect Technologies sells industrial inspection-as-a-service to Bharat Petroleum, Reliance, and Tata Steel; the sensors and the AI pipeline are wrapped in a contracted SLA. Unbox Robotics sells warehouse-throughput outcomes to logistics customers, not just robots.
The Indian industrial buyer rewards this. Listed-OEM procurement teams underwrite capability and uptime - they pay materially more for an integrated service than for a part. Founders who design from day one to own the operating contract, the data, and the recurring upgrade cycle build defensible decade-long franchises out of what would otherwise be a one-shot hardware sale.
The wave of Indian deep-tech founders arriving between 2024 and 2030 is not a single founding generation. It is the first of several. We are positioning to back two or three founding generations from inside the labs they are leaving.