LegendSpace, a six-month-old Beijing-based startup, raised $29.5 million in angel funding to develop liquid rocket engines using AI-driven design and simulation. Founded by Peking University aerospace engineer Chen Zhi, the company claims its "Prime" platform can compress engine development cycles from years to roughly two weeks by linking AI design generation, multiphysics simulation, additive manufacturing, and testing into a closed loop. The round reflects growing momentum in China's commercial space sector and follows the company's appearance in state media and a third-place finish at a global startup competition in July.
The funding will support multiple liquid engines through approval and testing, upgrades to the Prime platform to enhance AI simulation and metal additive manufacturing capabilities, and buildout of production infrastructure. LegendSpace has completed one full validation loop from AI-assisted design through simulation to both cold-flow and hot-fire testing, though the company has not specified what hardware was used. A 2.5-ton-thrust engine is currently in 3D-printed trial production, and the company is constructing a rocket engine test facility for static fire tests across various thrust levels.
The Prime platform builds on DeepFlame, an open-source GPU-native combustion simulation code from Peking University, and was launched in March as a collaboration between LegendSpace and the Beijing Institute for AI for Science. Chen's academic credentials include a Bernard Lewis Fellowship from the Combustion Institute in 2020 and prior work at Mitsubishi Heavy Industries, Sandia National Laboratories, and the German Aerospace Center. Chinese media coverage has drawn comparisons to SpaceX's use of AI in Raptor engine development.
LegendSpace enters a competitive market already occupied by dedicated commercial engine makers like Jiuzhou Yunjian and vertically integrated launch companies such as Landspace, Space Pioneer, and iSpace that develop propulsion in-house. Success will likely depend on demonstrating that AI-driven development can deliver measurably faster and cheaper engines at operationally relevant thrust levels. The company has not disclosed its engine's propellant combination, cycle, or prospective customers.