Critical Aerospace Closes 200 Million Yuan Round, Advancing AI for Science | Oasis Vitality

Oasis Capital led this funding round for LegendSpace.

Founded just six months ago, the team led by Peking University young professor Zhi Chen is attempting to reconstruct liquid rocket engine R&D through AI for Science and intelligent manufacturing — creating a closed loop from design, simulation, manufacturing to hot-fire testing, compressing development iterations from years to weeks.

As commercial space enters an era of high-frequency launches and scaled delivery, competition in propulsion systems hinges not merely on single-point performance, but on R&D efficiency, cost control, and mass production capability.

LegendSpace is tackling precisely the most fundamental — and most difficult — link in the industry chain.

LegendSpace (Beijing) Intelligent Technology Co., Ltd. (LegendSpace, hereinafter "LegendSpace") has officially completed a RMB 200 million angel funding round, with a consortium of top-tier VCs participating: Oasis Capital, Source Code Rhythm, Shangshi Capital, Qingliu Capital, Oriental Fortune Capital, and others co-invested.

This comes merely six months after LegendSpace's founding. Based in Beijing's Zhongguancun, this young company carries AI entrepreneurship in its DNA. It is a new-generation domestic liquid rocket engine R&D and manufacturer, steered by a 90s-born professor with an impressive resume — Zhi Chen. He graduated from Huazhong University of Science and Technology, then pursued doctoral studies in the United Kingdom, earning his PhD from the University of Cambridge and remaining there as faculty before returning to China full-time to join Peking University, where the prototype of LegendSpace was born.

It was only at the beginning of this year that Chen formally plunged into entrepreneurship, embarking on a path to upend traditional aerospace paradigms through AI for Science (AI4S). Through this moment, we witness an AI-driven revolution in aerospace manufacturing pulling back the curtain.

Founder Zhi Chen stated: "AI algorithms, numerical computation, and intelligent optimization are merely the means by which we deliver better engine products. LegendSpace's core positioning is as an aerospace propulsion equipment manufacturer. All our underlying technical breakthroughs and computing infrastructure build-out ultimately serve one purpose: producing cost-controllable, performance-leading, scalable self-developed rocket engines that fill the gap in domestic commercial space's independent propulsion supply."

Jinjian Zhang, Founding Partner of Oasis Capital, stated: "In the past, domestic commercial space capital focused predominantly on downstream rocket integrators and satellite constellation deployment, while high-quality targets in upstream core propulsion manufacturing remained extremely scarce. Overseas giants have built formidable moats through engineering accumulation; to break through, one must rely on new technological paradigms led by AI for Science and intelligent manufacturing. We believe the LegendSpace team combines both foundational theoretical innovation and engineering mass-production capability, capable of demonstrating lifecycle-spanning vitality."

From Cambridge Supervisor to Peking University Professor

A 90s-born academic betting on AI4S + Aerospace

Zhi Chen is undeniably the soul of LegendSpace.

Born in 1991 in Hunan, Chen's academic journey began at Huazhong University of Science and Technology, where he enrolled at age 17 in the Thermal and Power Engineering program. In 2010, he went to the University of Birmingham for mechanical engineering, earning dual bachelor's degrees from both institutions two years later.

After graduation, Chen continued to the University of Cambridge for his PhD in engineering. During this period, he built a multi-physics computational framework adapted to the extreme conditions of aerospace propulsion. Following his 2016 doctoral graduation, he successively conducted postdoctoral research and served as an independent senior researcher and doctoral supervisor at Cambridge's Department of Engineering and Robinson College.

For his outstanding research contributions in combustion theory and modeling, Chen received the Bernard Lewis Award from the International Combustion Institute in 2020. This prize is awarded every two years to no more than six recipients globally, representing the highest honor for young scholars in the international combustion field. Additionally, at the end of this month, Chen will be invited to deliver a plenary review lecture at the 41st International Combustion Symposium in Kyoto, Japan — the youngest and first-ever review speaker under 35 in the conference's century-long history.

Beyond his impressive academic achievements, Chen has a lesser-known record of top-tier international engineering experience: from 2016, he was deeply involved in tackling unstable combustion challenges in Mitsubishi Heavy Industries' new-generation heavy-duty gas turbines; in 2018, as a visiting scholar, he conducted exchanges at the Lockheed Martin-Sandia National Laboratories Combustion Research Facility in the United States; the following year, he served as a visiting scientist at the German Aerospace Center (DLR) Institute of Combustion Technology.

It was in 2021, inspired by DeepMind's series of breakthroughs — itself born from Cambridge — that Chen accurately foresaw AI's disruptive impact on scientific research. With full conviction, he decided to return to China to teach, joining Peking University's School of Mechanics and Engineering as an assistant professor, researcher, and doctoral supervisor. Simultaneously, from 2022 onward, he worked alongside Weinan E, the originator of the AI4S concept, and Linfeng Zhang, founder of DeepWise, to advance the establishment of the Beijing Institute of AI for Science, where he also serves as academic vice president. This institute is a new R&D institution deployed by Beijing for AI4S, and the world's first research institute named after AI for Science.

Notably, the Thermofluid Physics Intelligence Laboratory that Chen established at Peking University focuses on cracking the underlying methodology for forward design of propulsion systems, with numerous research outputs directly serving the R&D optimization of liquid engine combustion chambers, injectors, and thrust chamber core components. He also led the development of DeepFlame, the world's first open-source reaction flow simulation platform with a coupled AI framework, which won the "CCF HPC China 2025 Supercomputing Best Application Award." This award is among the most authoritative and influential in China's high-performance computing field, domestically regarded as a heavyweight honor comparable to the Gordon Bell Prize.

Over the past decade-plus, Chen has rooted himself in fundamental rocket propulsion mechanism research, moving across combustion fluid dynamics, propulsion structures, extreme-condition experimentation, and high-performance numerical computing. On this foundation, he has received numerous honors, the most eye-catching undoubtedly being the "Honorary Fellow of Robinson College, University of Cambridge" — making him the only Chinese lifetime Fellow in that college's nearly century-long history. Additionally, Chen was selected for the Royal Academy of Engineering's Exceptional Talent program at age 28, and was also the only young scholar in Peking University's engineering direction that year to be selected for the Central Organization Department's Young Top Talent Program.

Through years of observation and research, Chen discovered that the core bottleneck in domestic commercial rocket industry development remained concentrated in the propulsion end. Traditional engine R&D relies on massive physical iterative hot-fire testing, with lengthy development cycles and prohibitive prototyping and experimental costs. Pure trial-and-error can hardly support the industry demands of reusable rockets and high-density launch eras. "Our team's years of accumulated multi-physics coupling and intelligent optimization underlying research capabilities are fundamentally aimed at serving the efficient forward development of physical engine products."

Thus, at the beginning of 2026, Chen chose to step out of the laboratory and, together with his first doctoral graduate Runze Mao, founded LegendSpace, core-focused on AI-driven full-process self-developed liquid rocket engines, providing standardized, mass-producible propulsion system deliveries for commercial launch vehicle stages at all levels and aerospace vehicle power units.

At this March's Zhongguancun Forum, the "Prime" rocket engine R&D intelligent agent platform made its debut, jointly released by LegendSpace and the Beijing Institute of AI for Science, achieving a complete closed loop of "precise design — physical simulation — precision manufacturing — hot-fire verification," compressing the full-chain iteration cycle from rocket engine concept to physical verification to a "two-week" magnitude — whereas under traditional models this cycle is typically measured in years.

LegendSpace co-founder Runze Mao presenting at the event

This means the "Prime" platform not only avoids repeatedly "burning cash" on physical experiments, but transforms R&D from "trial-and-error with materials and luck" to "solving with data and algorithms." More importantly, this substantially lowers the technical threshold for domestic commercial space, with the bottleneck that has persisted at the propulsion end for years now poised for fundamental resolution.

2.5-ton thrust engine undergoing 3D printing

Top-Tier VCs Band Together for Heavy Bet

With its first funding round announced, LegendSpace has already made a deep impression on the venture capital circle.

This round's investor lineup consists entirely of domestic top-tier market-oriented VC institutions. Their concentrated appearance at the angel stage, jointly making heavy bets on the upstream core hardware manufacturing track in aerospace, is quite rare. As such, it sends a clear signal — not only confirming the long-term value of the AI + commercial space track, but also reflecting that primary market judgment logic is undergoing profound change.

From PEDaily's conversations, multiple investors participating in this round held convergent views. They generally believe that after more than a decade of development, China's commercial space industry has moved past the concept-hyping stage into a substantive implementation period of equipment mass production and high-frequency launches. As the core essential hardware of the launch system, rocket engines face continuously expanding market demand.

Meanwhile, the current global commercial space competitive landscape is forcing exponential growth in China's launch capacity needs and pressure. To catch up with or even surpass world-leading launch levels, one must jump out of established technical frameworks and achieve leapfrog overtaking by relying on new paradigms, new tools, and new systems. And hardware teams assembled by top young academic leaders, using AI to drive forward design and full-process lean manufacturing, represent precisely the most certain breakthrough path at present. Thus, LegendSpace appearing in investors' shared field of vision and becoming a consensus bet became an inevitable outcome.

As is widely known, compared to downstream and midstream segments of commercial space such as satellite applications and launch services, the upstream rocket propulsion hardware manufacturing gap is larger in terms of domestic production and intelligentization, with broader replacement space and more prominent long-term investment value. The primary market has clearly captured this industry trend and is voting with real money.

That LegendSpace could rapidly complete a RMB 200 million funding round right at its founding carries deeper significance: primary market judgment logic for early-stage commercial space projects has shifted from pure "sector positioning" to deep assessment of "technological paradigm reconstruction capability."

Along this logic, the AI for Science path that LegendSpace represents is precisely a typical target under this new evaluation framework: it is not simple replication of overseas mature routes, but fundamentally reconstructs rocket engine R&D methods from the bottom up.

The concentrated deployment of multiple top institutions at the angel stage is essentially a strategic bet on this "leapfrog overtaking" logic, also once again confirming an industry consensus — AI-empowered physical high-end manufacturing will become the core lever for China's commercial space to break technological blockades, narrow great-power competition gaps, and achieve leapfrog development.

To some extent, this funding round will establish a benchmark for early-stage commercial space projects, also heralding that a new investment era of AI-empowered aerospace manufacturing may be arriving.

China's Commercial Space Leapfrog Overtaking

At present, China's commercial space is approaching an inflection point.

From an industry perspective, commercial space's strategic positioning is undergoing a historic leap. Recall late last year, when the Shanghai Stock Exchange announced that commercial rocket enterprises would be eligible for the STAR Market's fifth set of listing standards, stunning the venture capital circle; entering this year, commercial space was formally incorporated into the "15th Five-Year Plan" key tasks, with a dedicated commercial space regulatory bureau established. Such measures construct a complete system spanning top-level planning, industry regulation, and capital markets. The Long March 10B carrier rocket also successfully completed its maiden flight, validating the world's first "sea-based net recovery" technology.

Meanwhile, demand-side growth is exploding. Across the country, multiple localities are accelerating aerospace industry cluster deployment, competing to form complete industrial ecosystems integrating R&D, manufacturing, testing, and application; batches of reusable launch vehicles and low-orbit satellite constellation projects are also being intensively advanced.

This rapidly expanding market demand places new requirements on rocket propulsion systems. Traditional high-cost, long-cycle, low-reusability propulsion products can no longer adapt to scaled commercial launch scenarios; what the market increasingly needs is high-reliability, low-cost commercial space launch capacity and propulsion systems.

For LegendSpace, this is a historical window that cannot be missed.

Zooming out to the global picture, we can see that overseas leading aerospace enterprises have long completed deep integration of AI and aerospace manufacturing, relying on intelligent systems to achieve rapid iteration, extreme cost reduction, and scaled mass production of propulsion products, continuously consolidating technological leadership advantages.

Take SpaceX, which recently completed the largest IPO in history. Beyond leading the industry in rocket recovery technology, it also extensively applies AI technology in engine design and manufacturing. Materials show that SpaceX, by integrating AI tools from the initial design phase, substantially shortened development time and material costs, achieving rapid iteration without massive physical testing. Its Raptor 3 engine has nearly doubled thrust while costs dropped to one-quarter of the original — this is precisely the most intuitive manifestation of an "AI-driven manufacturing revolution."

SpaceX's practice proves that for domestic enterprises, simply replicating overseas traditional R&D paths offers no breakthrough possibility. AI intelligent manufacturing is already the core trend in commercial space propulsion industry upgrading, and the key to breaking technological bottlenecks and achieving scaled commercialization. Only by relying on the new paradigm of AI intelligent manufacturing to reconstruct the full-process system of rocket propulsion R&D, manufacturing, and testing can we break years of technological generation gaps and achieve transition from passive chasing to active surpassing.

Chen feels this deeply. He judges that the space economy is a long-term track with extremely high certainty over the coming decades, while propulsion equipment is the foundation of entire industry development. The barrier of overseas aerospace propulsion is essentially a barrier of accumulated traditional engineering iteration; latecomers find it difficult to catch up linearly.

"AI intelligent manufacturing is a brand-new opportunity belonging to China's commercial space. Our long-term goal is to build a domestic-leading, globally-competitive independent liquid rocket engine R&D and manufacturing platform, using the new paradigm of intelligentization to polish high cost-performance, mass-producible domestic rocket engines, helping domestic commercial space achieve leapfrog overtaking, and supporting major industry projects such as low-orbit constellation deployment and deep space exploration."

It is understood that the "Prime" platform has completed its first successful closed-loop case verification from AI-assisted design, through simulation evaluation, to cold-flow/hot-fire test validation. The entire process is agent-driven, including multiple sub-modules of design generation, parametric modeling, simulation evaluation, multi-objective optimization, and test planning and data feedback, working autonomously and collaboratively, connecting candidate configuration generation, physical evaluation, test verification, and error feedback in series to ultimately form a closed loop.

This platform not only supports cutting-edge research directions including AI generative design, AI-driven combustion simulation, metal additive manufacturing, and automated fluid measurement and control, but also possesses autonomous iteration and evolution capability — capable of precisely screening and locking in optimal material solutions from vast elemental permutations.

PEDaily also learned that following this funding round, LegendSpace will accelerate the finalization and experimental hot-fire testing of multiple engines at different thrust levels, while simultaneously engaging more aerospace industry chain industrial capital, building capital reserves for subsequent production delivery.

This also marks that current primary market capital will continue tilting toward hard manufacturing enterprises led by scientists who possess physical products, complete engineering implementation capability, and mastery of new paradigm core technologies.

It is foreseeable that a wave of domestic aerospace propulsion revolution driven by AI intelligent manufacturing is accelerating. From "following" to "running alongside" to "leading," the critical point for China's commercial space may be right before our eyes.

What is vitality?

For entrepreneurs, vitality is the unprecedented power that emerges from activating the hidden parameters of life's large model. — Zhi Chen, Founder of LegendSpace