Agnikul Cosmos Clusters Four 3D-Printed Semi-Cryogenic Engines in Record Static Fire

Indian First The Problem: Making Four Engines Behave as One

Agnikul Cosmos has fired four semi-cryogenic rocket engines simultaneously, all 3D printed as single-piece components and developed entirely at its Rocket Factory 1 facility in Chennai. The company claims this is the first clustered semi-cryogenic engine test in India.

The real challenge wasn't the engine count. It was synchronization. Each engine runs two electric pumps, each pump has its own motor, and each motor runs its own speed control algorithm. Eight pumps, eight motors, eight algorithms, all needing to start together, hold steady together, and shut down together. Get it wrong and you have asymmetric thrust, which in a flight vehicle means control authority problems or structural loads you didn't design for.

> "This test involved calibrating eight pumps, eight motors and tuning eight speed control algorithms to work together in perfect sync to achieve uniform startup, steady state and shutdown performance across the entire system," said Srinath Ravichandran, Co-founder and CEO of Agnikul Cosmos. The Architecture: Why Electric Pumps, Why Now

Agnikul's engines use electric pump-fed propulsion instead of conventional turbopumps. The trade-off is straightforward: turbopumps give you higher performance per kilogram of propellant, but they're expensive to manufacture, hard to throttle, and a nightmare to refurbish between flights. Electric pumps sacrifice some of that performance for operational simplicity. You vary motor speed to change thrust, which gives you finer throttle control and fewer parts that need replacing.

Rocket Lab proved the concept with its Rutherford engine, the first 3D-printed, electric pump-fed engine to reach orbit. Rutherford has now flown on more than 70 Electron missions, with over 1,000 units produced. New Frontier Aerospace is pushing the architecture further with its Mjölnir engine, a full-flow staged combustion design targeting hypersonic VTOL in 2026 and an orbital transfer vehicle in 2027.

The unsolved problem Agnikul is attacking is cluster synchronization. Single-engine electric pump tests are now routine. Getting multiple engines to run in phase with each other is a different class of problem. The Manufacturing Story: Days, Not Months

Every engine in this test was printed in days, designed and built entirely within Rocket Factory 1 with no external tooling or fabrication partners. Agnikul's business model depends on this. Their roadmap assumes production throughput matches customer demand, and conventional fabrication timelines would break that assumption.

This four-engine test follows a quickening pace of milestones. In December 2025, Agnikul completed what it described as India's first dual pump-fed engine firing, two engines running together for 49 seconds. In March 2026, the company ran a full design-to-test cycle on the Agnite, a one-meter-long single-piece Inconel semi-cryogenic engine, at its Chennai test facility. What Happens Next

Clustered electric pump propulsion at flight scale remains unproven for Agnikul. The test validates the control architecture and manufacturing approach, but flight environments add vibration, thermal cycling, and the structural loads of actual ascent. If the synchronization holds under those conditions, Agnikul joins a very short list of companies that have built clustered electric pump systems from scratch.

The turbopump isn't dead. It still wins on specific impulse and mass ratio. But for small launch providers optimizing for cost per flight and turnaround time, the electric pump architecture is looking less like a compromise and more like a deliberate design choice. Agnikul's test pushes that choice from the single-engine demonstration phase into something that could actually fly.

M4S TAKE

My take: capacity expansions signal confidence, but the real question is whether demand justifies the spend. I watch for follow-up announcements about utilization rates or new contracts. Without those, this is just capital allocation.

Simon McLoughlin