Agnikul Cosmos is a private Indian launch company based in Chennai that builds small rockets around a single-piece, 3D-printed semi-cryogenic engine. Incubated at IIT Madras and founded in 2017, it flew the world’s first rocket powered by a fully 3D-printed engine in May 2024 and has since concentrated on the harder problem of clustering several of those engines into a first stage. It has not yet reached orbit.[1][2]
Origins
Agnikul began in 2017 as a bet on manufacturing rather than on rocket science. Srinath Ravichandran, an aerospace engineer, and Moin SPM, a mechanical engineer, teamed up with S. R. Chakravarthy, a professor at IIT Madras, around a simple premise: if a rocket engine could be printed as one piece instead of assembled from hundreds of machined parts, the cost and lead time of access to space would fall sharply. The company is incubated at IIT Madras Research Park in Chennai, and its early work was supported by the institute’s propulsion and additive manufacturing expertise.[3][4]
The engine that made the company
The Agnilet is the centre of everything Agnikul does. It is a semi-cryogenic engine — burning liquid oxygen with aviation turbine fuel — printed as a single piece of nickel superalloy, with no welding or bolted joints, and fed by electric motor-driven pumps instead of turbopumps. Each engine produces about 6.2 kN of thrust, stands roughly a metre tall, and is printed in-house. The company says a complete engine can be turned around in about a week, against the months a conventionally assembled engine takes.[5][6]
A larger engine, Agnite, extends the same approach to the booster stage. Also printed in one piece and also semi-cryogenic, it was test-fired in 2026 and is described by the company as India’s first semi-cryogenic engine and the largest single-piece 3D-printed rocket engine in the world. In October 2026 it completed multiple thrust sweeps and a long-duration firing exceeding the burn time of some of the company’s ascent-stage engines — the qualification work needed before it can power a flight.[7][8]
Printing an engine in one piece is not a manufacturing trick; it is a business model. It moves the constraint from how many parts you can source to how many engines you can print, and that is what a launch cadence is actually made of.
SOrTeD and the launchpad
The company’s first flight was a suborbital demonstrator, and it took several attempts to get away. Agnikul built India’s first private launchpad and mission control centre at Sriharikota, inaugurated in November 2022, and scheduled attempts in March, April and May 2024 were called off — one of them 92 seconds before lift-off. On 30 May 2024, Agnibaan SOrTeD finally flew from Agnikul Launchpad-01 at 07:15 IST, executing a clean lift-off, a pitch-over and wind-biasing manoeuvres during a 66-second powered flight before coming down in the Bay of Bengal.[9][10]
The flight packed several firsts into under a minute and a half: the first launch anywhere of a rocket powered by a single-piece 3D-printed engine, India’s first flight of a semi-cryogenic engine, and the first Indian launch from a privately built and operated pad. SOrTeD was also the first private Indian vehicle to use ISRO’s flight termination system packages, and the demonstration validated the company’s guidance, navigation and control software as much as its propulsion.[11][12]
Scaling from one engine to a cluster
Agnibaan is designed to be configured rather than fixed. Its first stage can carry anywhere from four to seven Agnilet engines depending on the payload, and the company’s pitch is that a customer buys the rocket their satellite needs rather than paying for excess capacity on a standard vehicle. The full vehicle is about 18 m long and 1.3 m in diameter, with a lift-off mass of roughly 14,000 kg and a capacity of up to about 300 kg to low Earth orbit.[13][14]
That modularity is only worth anything if the engines fire together, and 2026 was spent proving they do. The company worked up from two engines to three in February and to four in May, the four-engine test requiring the avionics to coordinate eight pumps, eight motors and eight speed-control algorithms for uniform start-up, steady-state operation and shutdown. All four were single-piece 3D-printed engines made in-house. The remaining steps are stage-level testing — firing the full complement in a ground-clamped configuration — and then the stage separation that every clustered first stage has to get right.[15][16]
Mission-02 and the reusability bet
Agnikul’s next flight is not simply an orbital attempt; it is an attempt to bring hardware back. Mission-02, using a two-stage Agnibaan, is planned to put customer satellites in orbit and then fly the first stage back for a powered vertical landing on a barge in the Bay of Bengal, south of Sriharikota. The booster will carry landing legs, grid fins and an independent flight computer so it can guide itself home after separation. The upper stage, meanwhile, is meant to stay alive in orbit as a convertible platform rather than becoming debris — the company holds a patent on extending upper-stage life.[17][18]
The ambition has now been backed with public money. In September 2026 the Technology Development Board of the Department of Science and Technology signed an agreement for âš200 crore of support, through optionally convertible debentures, to develop a fully reusable Agnibaan RLV — aiming to move the technology from TRL-4 and above to TRL-8, with a lightweight upper stage, deep-throttling and restartable semi-cryogenic propulsion, and descent and recovery systems. Notably, the programme targets full-system reusability rather than recovery of the first stage alone.[19]
Money, factory and cadence
Agnikul has raised more than $75 million in disclosed equity, including a âš150 crore ($17 million) round in November 2025 that valued it at $500 million, backed by Advenza Global, HDFC Bank, Artha Select Fund, Prathithi Ventures, 100X.VC and family offices. In March 2026 Tamil Nadu’s TIDCO invested âš25 crore directly, a rare case of a state government taking equity in a private space company[20], and reports in May 2026 described a further $50–75 million being raised at a flat $500 million valuation.[21][22]
The money is going into a 350-acre integrated space campus on land allotted by the Tamil Nadu government near the coming Kulasekarapattinam spaceport in Thoothukudi district, chosen for its proximity to the new launch pad. Built in phases over two to three years, it is intended to produce 25 to 30 rockets a year when complete. Agnikul has spoken of 50 launches a year by 2028 — a figure that will be judged against a company that reported a âš43 crore loss on âš9 crore of revenue in its most recent disclosed financial year.[23][24]
The orbital gap
Agnikul’s most awkward fact is comparative. Rival Skyroot Aerospace reached orbit on its first attempt in July 2026, and ISRO has flown its own small launcher, so Agnikul is now chasing a milestone that has already been claimed twice over in India. Its own orbital attempt has slipped from an early-2026 target to late 2026 or the first half of 2027, with missions on its published launch schedule sized for up to 100 kg and then up to 300 kg to low Earth orbit.[25]
What it is betting on is a different prize: a rocket cheap enough to print, configurable enough to sell by the kilogram, and reusable enough to fly again. That is a harder engineering programme than a single successful orbital debut, and a more valuable one if it works. Whether it does will be settled by two events — the first full orbital flight, and whether the booster comes back — neither of which has happened yet.[26]
Select a year and month above to explore cold archived discussions.
Please select an archived topic above to view the discussion logs.