Musk’s Starship was launched again. This time, all 33 Raptor 2 engines of the first-stage super-heavy booster were successfully ignited. The spacecraft completed a thermal startup mid-air and reached an altitude of 145 kilometers, surpassing the Kármán line into space, but failed to enter the intended orbit. This marks a significant improvement compared to the April launch, which only reached 30 kilometers.

01

Every time SpaceX conducts a major launch, it becomes a celebration for space enthusiasts. However, I also notice two other groups of people: one revels in American invincibility, while the other shows irritation, anxiety, and even rage, seizing any chance to criticize Musk and SpaceX. In any ordinary event, I’d consider these two groups to be just spouting clichés for the sake of argument, but SpaceX’s case is somewhat different. If Starship successfully launches, it signifies a step closer to becoming a spacefaring civilization or colonizing Mars and that in this new space race, post the US-Soviet rivalry, our gap with the U.S. has genuinely widened. The stakes of this matter go far beyond mere rhetoric.

The first-stage super-heavy booster of Starship stands at 70 meters, with the spacecraft at 50 meters, making the total height around 120 meters. This surpasses the Saturn V, the previous largest rocket used for manned lunar missions, by 10 meters. It towers over China’s largest rocket, the Long March 5, and SpaceX’s operating Falcon 9; the first stage of Starship contains 33 Raptor-2 engines, each with a thrust of 230 tons, amounting to a total of 7590 tons. This is more than double that of Saturn V’s 3408 tons and substantially exceeds Falcon 9 and Long March 5B. When reusable, Starship can carry 150 tons to low Earth orbit and 100 tons to the Moon or Mars, surpassing the single-use Saturn V and being about six times that of Falcon 9 and Long March 5B.

This means that as Starship matures through iterations, it will inevitably greatly enhance SpaceX’s space transport capabilities and reduce transportation costs.

02

As of this year, there have been 187 rocket launches globally, with Chinese organizations responsible for 53 and SpaceX alone for 85. Except for two Starship launches, all were successful. SpaceX’s cost per kilogram to near-Earth orbit is now less than 10,000 RMB, roughly one-twentieth of China’s corresponding cost. Emphasizing this point again, the cost of our launches is 20 times that of SpaceX.

If the US-Soviet space race was ostensibly a display of national power but actually centered around military competition, serving as a reminder of their long-range strike and comprehensive surveillance capabilities, then what is the theme of this current space race in an era where intercontinental missiles and remote sensing satellites are already mature?

On August 29 this year, during U.S. Commerce Secretary Raimondo’s visit to China, Huawei suddenly released its new flagship phone, Mate60pro, with its most unique feature being the capability to make satellite calls. This can be lifesaving in remote areas and provides a fallback for high-profile individuals concerned about eavesdropping through telecom operators. Following this, on October 11, SpaceX’s Starlink announced plans to offer direct satellite connectivity to mobile phones, with text messaging expected by 2024 and voice calls and internet access by 2025. Eventually, even IoT devices like rice cookers and automatic cat litter boxes could connect to Starlink. The service does not require a special phone, just 4G compatibility.

Subsequently, on November 3, Musk announced on Twitter that his Starlink project had achieved breakeven. The industry estimated this six months ago, with SpaceX’s satellite launch and ground operations costing around $2.4 billion. Selling dishes to 1.5 million users and collecting subscriptions roughly equals this amount. The fact that so many people need internet in remote areas was beyond my imagination. Plus, Starlink’s well-known support to the Ukrainian military, which lost ground networks in the Russia-Ukraine conflict.

This couldn’t be clearer: humanity is entering the era of the constellation internet. Since the transition from 2G to digital networks, our mobile networks have been called ‘cellular networks,’ meaning they use a honeycomb-like structure of base stations to cover large areas.

However, this coverage is always limited and costly. Additionally, higher frequency electromagnetic waves, while carrying more information, have weaker diffraction ability. As mobile networks evolved from 2G to 3G, 4G, and 5G, base stations became increasingly dense, driving up costs. No matter how much a country invests in infrastructure, it’s a bottomless pit. Cellular networks can only cover densely populated areas.

So, the good news is once satellite networks cover the Earth, an era of uninterrupted internet connection everywhere and at all times, with no dead zones, will finally arrive. Hikers, camping enthusiasts, R.V. fans, off-roaders, astronomers, wildlife photographers, drone enthusiasts, sailors, and pilots will all be overjoyed.

But the bad news is, according to Shannon’s theorem, a channel’s maximum theoretical transmission rate is positively correlated with bandwidth and signal-to-noise ratio. To improve the signal-to-noise ratio or signal strength, satellites dedicated to the constellation Internet must be placed in low Earth orbit.

But space in low Earth orbit is limited. If satellites are placed too densely, an explosion of one satellite could trigger a chain reaction, potentially blocking the entire low Earth orbit with debris. Some estimate that Earth’s low Earth orbit can only accommodate about 60,000 satellites.

Internet communication satellites, compared to previous research, navigation, and remote sensing satellites, are characterized by their large numbers due to the massive data transmission volume. The Starlink project alone, once fully launched, will take up 40,000 of that quota, and SpaceX has already launched over 5,400 satellites.

More troubling is that you cannot verify a satellite’s functions when it’s in space. Starlink claims to be a communication satellite, but SpaceX has a project with the U.S. Department of Defense for the Starshield reconnaissance satellite. Eventually, 40,000 reconnaissance satellites could feasibly cover the globe without dead zones. Previously, reconnaissance satellites only focused on collecting data from key facilities because there were too many images to process. But with the A.I. revolution, this is no longer a problem, allowing for precise targeting of a tank, an airplane, or even a person.

‘K’ with abundant martial virtue, high danger level, prioritize elimination;

‘M,’ addicted to Genshin Impact, danger level 0, ignore.

This is quite distressing.

03

So, how do we compete for space in low Earth orbit with an entity whose launch costs are just a twentieth of ours and are continually decreasing? The only way is to develop our reusable rocket technology, increase capacity, and lower costs. On November 2 this year, a private rocket company called i-Space successfully conducted a first-stage recovery test of their Hyperbola-2 rocket at Jiuquan Satellite Launch Center, marking the most advanced progress in China’s reusable rocket technology.

To understand the level of this progress, let’s review SpaceX’s development journey in reusable rockets. If you’re not a deep space enthusiast, you might have the impression that SpaceX rockets frequently explode. While researching this article, I asked several relatives and friends who don’t follow space news, and the most common misunderstanding is that Musk, a private entrepreneur, reduces costs by cutting corners, leading to failures. If that were true, it would imply that SpaceX rockets lack technical sophistication and that we could reduce costs by simplifying the design.

Unfortunately, that’s not the case. All rocket models explode during development – engines explode during ground tests, pressure tests of the hull might explode, and launches might explode. This is true for rockets worldwide. Otherwise, do you already know all possible anomalies? Are you a time traveler? The strange thing about SpaceX’s rockets is that even after a model has successfully launched, it often still explodes. This differs from other rockets, where once a model is successfully launched, it’s set and not modified further.

Because what are rockets for? The primary purpose of rockets is to deliver a specific weight to a specific orbit. And what does launch success mean? It means delivering a specific weight to a specific orbit. Since it’s already been achieved, why change anything? Any subsequent modification only adds unnecessary risk factors without any benefits. However, SpaceX’s rockets were developed with a different goal. Musk aims to transform humanity into a spacefaring civilization, to colonize Mars, and more precisely, to outcompete all global rocket launch institutions cost-effectively, counting his small money while reducing the cost of rocket transportation to an extreme degree, ultimately targeting ten dollars per kilogram. Therefore, he does not stop iterating development just because of launch success.

The Falcon 9, currently providing humans with the strongest space transport capability, had its first successful launch in June 2010. Usually, this would be the time to finalize the model, but it wasn’t until December 2015 that Falcon 9 achieved its first recovery of the first stage after nine failed recovery attempts. They first tried to recover the first stage with parachutes, which failed. Then, they attempted to slow down with a spectacular secondary ignition upon re-entry and tried both land and sea recovery. Although this story has been told many times, I still find the recovery footage of Falcon 9 thrilling – it’s unbelievable.

But did you think the recovery marked the end? By that standard, wouldn’t the Space Shuttle, which is also recoverable, suffice? The Space Shuttle had 135 launches, two of which exploded; out of the five built, two were destroyed, killing 14 astronauts. One notable disaster was the Challenger in 1986, which carried a high school teacher preparing to give a lesson but ended up giving everyone a harsh lesson instead. The average launch cost of the Shuttle was over 450 million dollars, with a per-kilogram transport cost of over 50,000 dollars, making it the most expensive and unsafe spacecraft.

The difference lies in the Shuttle’s overly complex refurbishment after recovery, requiring 650,000 hours of labor and prone to errors. In contrast, SpaceX aims to refuel and relaunch immediately after recovery. Therefore, after the first successful recovery on December 12, 2015, Falcon 9 underwent numerous iterations, including upgrading the Merlin engine’s turbopump due to microcracks found in previous recovered engines, changing the grid fins controlling direction from aluminum to more heat-resistant titanium, adding a thermal coating to the hull, adding a liquid-cooled heat shield at the engine, making the landing gear retractable, upgrading the helium coolant pipelines in the fuel tank, and more. Only after the launch of Falcon 9 Block 5 on May 11, 2018, did it truly achieve reusability, with four more explosions in between. The most reused Falcon 9 has launched 18 times, with only four new Falcon 9 rockets used in 2023’s 83 launches. It’s estimated that a Falcon 9 can be launched 100 times.

To ultimately achieve immediate relaunch after recovery, SpaceX continues to fundamentally reform its rockets. For instance, the Raptor 2 engine used in Starship, compared to the Merlin engine used in Falcon 9, has a notable difference: it uses liquid oxygen and methane instead of kerosene. Kerosene inevitably leads to carbon buildup, requiring engine disassembly and cleaning before each reuse, a time-consuming and laborious process. Methane, or natural gas, is a much cleaner fuel and addresses this issue at its root.

04

Do you think it’s over just because the rocket launches quickly? Far from it; we need to launch a lot more! Since rockets are now being made reusable, a portion of the fuel must be reserved for deceleration during re-entry into the atmosphere and for the final braking. This will consume some of our already limited payload capacity. Although the rocket launches quickly, it carries a cute but small payload, which is not ideal. To ensure that a reusable rocket can carry a payload equal to or even greater than traditional rockets, we need to develop engines with a higher thrust-to-weight ratio.

The volume of the rocket is limited, and to store as much fuel and oxidizer as possible, they must be cooled into liquid form. The boiling point of oxygen is minus 183 degrees Celsius, and methane’s is minus 161 degrees. However, these two low-temperature liquids cannot mix perfectly in the combustion chamber, limiting combustion efficiency and thus thrust. The Raptor 2 engine adopts an unprecedented structure, allowing liquid methane and liquid oxygen to slightly mix and ignite in two small combustion chambers before entering the main chamber. They are then heated and transformed into gas and enter the combustion chamber at high speed, increasing the combustion efficiency to an astonishing 99%. During its development, at least 50 engines were destroyed.

The Starship is equipped with 39 Raptor 2 engines, each with a thrust of 230 tons, far surpassing the previous generation Raptor 1’s 185 tons and the Space Shuttle’s RS25 engine’s 190 tons. The Raptor 2 weighs only 1600 kilograms, while the RS25 weighs 3200 kilograms.

Even this is not enough. After achieving his targeted results, Musk’s next goal is to simplify the production process and ultimately achieve large-scale automated production, significantly increasing output and reducing costs.

You don’t even need to understand the engine’s structure to see how much simpler Raptor 2 is compared to Raptor 1. Last year, SpaceX could already produce one Raptor engine per day.

For example, the Falcon Heavy, in addition to its main body, has two booster rockets. This design isn’t unique, but why do these boosters look so familiar? Because they’re the same as the first stage of the rocket itself. Why open a new production line when one can solve the problem?

This rapid iterative development approach based on extreme goals is far from shoddy. To make this development method work, each rocket must have high consistency first. Otherwise, if one explodes and the issue is addressed, the next might explode differently, making it impossible to manage. Second, they must quickly and accurately identify problems to maximize the iteration speed. Although only seven months separated Starship’s two test launches, with a delay by the FAA, SpaceX rapidly implemented over 1000 improvements, backed by advanced sensors, data links, and digital twin technology.

This is difficult, but once this development process runs smoothly, your rockets will not only be the most efficient and cost-effective in the world but also the safest. To date, Falcon 9 Block 5 has had 223 launches with a failure rate of zero.

Hence, we understand why they sometimes applaud even after an explosion. Because real progress is being made, last time it could only fly 100 meters; this time, it reached 30 kilometers, and I know the issue will be quickly identified, making the next launch better than the last. Isn’t applauding natural in this case?

If we want to win this space race, we, too, need to repeat this iterative development process. This means in the coming years, we will see Chinese rockets also exploding non-stop. First during test runs, then trying to deliver payloads to the designated orbits, exploding, then attempting controlled re-entry and landing, exploding, and so on until success, followed by continued iterations and explosions until we approach or even surpass Falcon 9’s cost level.

As for how many years this process will take, in September 2020, China StarNet filed with the ITU for two low Earth orbit satellite constellations totaling 12,992 satellites, which must be launched within 7 years, or the rights will be relinquished. Assuming each launch can carry 30 satellites, 433 launches will be required to complete the task. According to industry experts, mass production and launching of these rockets will take at least two years. Therefore, we must acquire mature, mass-produced reusable rocket technology by September 2025, leaving us with less than two years.

There’s not much time left for the Chinese team.

Conclusion

This is undoubtedly a huge challenge. Although we don’t need to worry about China’s basic industrial capabilities or the determination to compete, the timeframe is indeed very tight. Previous hesitations in developing reusable rockets have been somewhat of a delay. But this is also a tremendous opportunity because if a Chinese company successfully develops reusable rockets, the first wave of orders will be for at least ten thousand satellite launches. And since communication satellites have an operational life of only about 5 years, a constellation of ten thousand satellites corresponds to a replenishment demand of 2000 satellites per year. So, I’m closely studying this matter. Next, I plan to observe more launches and delve deeper into this issue. Of course, while we have confidence in China’s speed, we should not forget a more fundamental question. Disruptive innovation has once again occurred in the U.S., in the West, making us followers again. This does not fit our status. Next time, how can we make it happen in China?