Foreword:

The performance boundaries of a chip are quietly shifting from transistors to the connections between chips.

In 2026, a fascinating disjunction has emerged in China's private rocket sector: the competition for medium-sized rockets is far from over, yet the gunshots in the 100-ton category have already begun to ring out Intense ringing .

Author | Fang Wensan

Image source | Internet

Private Rocket Sector Collectively Rushes Toward 100-Ton Class

i-Space announced on its 10th anniversary that pre-research has begun on its heavy-lift model, the "Hyperbola-9" rocket. With a takeoff thrust of 7,000 tons and a near-Earth orbit capacity of 100 tons, it features 33 parallel 300-ton liquid oxygen-methane engines. On the same day, Hengzhou Yuyuan, founded in 2026, included a 100-ton reusable rocket in its development plans.

This is not an isolated event. CAS Space is planning the 40-ton near-Earth orbit "Lijian-3," while Orient Space has already put the "Gravity-3" on the table before the maiden flight of the "Gravity-2." Galactic Energy, following the successful orbit insertion of the "Pallas-1," is betting on the 20-ton "Pallas-2."

If we consider only the urgency on the demand side, the answer should not be the 100-ton class. The first phase of the Qianfan constellation plans for 648 satellites, with a target of 324 in orbit by the end of 2026 to form initial service capabilities. The GW constellation needs to deploy about 10% of its 1,300 satellites by the end of 2029; otherwise, it risks losing frequency and orbital resources. As of mid-September 2026, approximately 256 Qianfan satellites are in orbit, about 68 short of the year-end target. Zhang Xiaodong, chief designer of Landspace's Zhuque-3, and his team calculated that to deploy over 20,000 satellites, China will need an average of 500 medium-to-heavy launch vehicles annually over the next 7 to 10 years. However, in 2025, only a small fraction of launch missions met the medium-to-heavy definition.

What truly constrains constellation networking is stable, schedulable, and cost-controlled medium-sized capacity, not the 100-ton class. Among the launch missions completed by the Qianfan constellation in 2026, the Long March 8 and Long March 6A accounted for 80%. The gap in private rockets lies first and foremost in the medium-sized category.

So why are leading companies skipping ahead? The first motive is the capital narrative. In the context of STAR Market IPOs and primary market financing, labels like "benchmarking Starship," "supporting deep space exploration," and "Earth-Moon space missions" unlock far greater valuation potential than a medium-sized rocket with a 7-ton capacity. i-Space's Series E funding plan aims to raise 7 billion yuan, a level of capital demand that requires a sufficiently grand narrative to support.

The second motive is "entry ticket anxiety." The technological spillover effects of 100-ton heavy-lift rockets are real: 300-ton engines, fault isolation for 33 parallel engines, and full two-stage reusability. Once these capabilities are achieved, downward compatibility with medium-sized capacity is a natural progression. Conversely, climbing up from medium-sized rockets requires rebuilding the development system at every step. Private companies are choosing to bet on the highest difficulty during the capital window, gambling on a "one-leap, full-spectrum" approach.

The third motive is the urgency created by SpaceX's Starship. The Starship V3 has a single-launch near-Earth orbit capacity exceeding 100 tons, with 33 Raptor 3 engines providing a total thrust of 9,240 tons. It achieved orbit for the first time in its 14th test flight. When the benchmark has made the 100-ton class routine, Chinese private space companies risk being skipped over in the next round of industrial discourse if they remain focused on "medium-sized rocket competition."

The '100-Ton Class' Masks the Real Barriers

As of June 2026, China's commercial space sector had 88 liquid rocket engines in development or service, with 36 falling in the 80-to-150-ton thrust range, accounting for about 58%. The 100-ton class has become the most crowded thrust range for domestic commercial rocket engines.

However, thrust figures alone are the easiest indicator to mislead. An engine that has completed multi-engine parallelization and entered actual flight can be called "100-ton class," as can an engine that has just completed its first full-system test. A gas-generator cycle engine with extensive flight validation can share the same parameter table as a more aggressive full-flow staged combustion engine still in ground testing. The Merlin 1D's cycle is not novel, yet it supports Falcon 9 in establishing the world's most mature orbital-class reusability system. What is truly scarce is not "building a 100-ton engine" but "a 100-ton engine that can be stably delivered, continuously flight-validated, and ultimately support recovery and reuse."

An even tougher hurdle than engine parameters is system integration. Parallelizing 33 engines is not difficult; the challenge lies in identifying and isolating single-engine failures. The 14th Starship test flight provided a concrete example: when a vacuum-optimized Raptor on the upper stage shut down prematurely, the flight control team needed to determine whether the faulty engine affected the propellant supply to the remaining engines. Only after confirming "mutual isolation" did they decide to proceed with orbital insertion ignition. This judgment relied on real-time data, which SpaceX had accumulated over 14 test flights.

From "orbital insertion" to "recovery" to "reuse," each step has independent failure modes. Landspace's Zhuque-3 successfully reached orbit in its maiden flight in December 2025, but its first stage experienced an anomaly during landing segment ignition. Land recovery was not achieved until August 19, 2026, with a near-Earth orbit capacity of 18.3 tons in recovery mode. China's reusable rockets are still "far from engineering application."

The Tianlong-3, with a near-Earth orbit capacity of 17 to 22 tons, is one of China's most powerful commercial liquid rockets. However, it experienced an anomaly during its maiden flight on April 3, 2026, and the mission failed. Before the gunshots in the 100-ton category, the competitive edge of medium-sized rockets had not even been fully tested.

How Many Can the Market Sustain After the Gunshots?

As of 2026, China has 50 companies capable of overall design and development of launch vehicles, with about 70 rocket models in development or service. Thirty-seven rocket factories plan an annual capacity of 610 launches, while 22 liquid rocket engine factories disclose a combined annual capacity of about 3,560 engines. The supply-side numbers are already substantial, yet the demand-side structural characteristics are unfavorable to "skipping levels": the most urgent needs for constellation networking currently focus on single-launch medium-sized capacities of 10 to 20 tons. The applications for 100-ton heavy-lift rockets—large-scale space infrastructure, deep space exploration, and Earth-Moon transportation—are far from entering Intensive procurement cycle a dense procurement cycle between 2026 and 2028.

Capital has already begun to view this with a more sober framework. A common judgment in the industry is that only two to three survivors are likely in this round of heavy-lift rocket competition. Companies betting on skipping to the 100-ton class are essentially trading their current market share in medium-sized capacity for entry tickets to tomorrow's heavy-lift market. Whether this trade-off is worthwhile depends on two variables: whether the capacity gap for constellation networking can first translate into revenue for medium-sized rockets, and whether heavy-lift rockets can achieve technical milestones—maiden flight and recovery—before cash flow runs out.

Conclusion

The competition for medium-sized rockets remains undecided, yet the gunshots in the 100-ton category have already sounded. This is because the race is not even about technology but about who can turn the "100-ton" label from a thrust specification into a flight resume before capital patience runs out.

Online References:

TMTPost: "Stripping Away the '100-Ton Class' Label: What Remains for Commercial Rocket Engines?"

East Money: "Private Commercial Space Companies Challenge Heavy Reusable Rockets, Deep Space Exploration Sector Heats Up"

36kr: "Launching While Listing: What Are Commercial Space Companies Competing For?"