After burning through money for several years without getting on track, what on earth is Elon Musk's "Starship" up to?

At 7:23 Beijing Time today, a giant rocket standing 123 meters tall soared into the sky in Boca Chica, a coastal town in Texas, USA.

This is the "Starship" being developed by SpaceX, the largest and most powerful rocket in human history, and also the key tool for Elon Musk to realize his dream of reaching Mars.

The steel behemoth let out a thunderous roar in the setting sun, shattering the tranquility of the beach. This was the eleventh orbital launch of the Starship, the fifth this year, and also the swan song of the second - generation Starship that SpaceX had high hopes for.

The Swan Song of the Second - Generation Starship

In terms of the result, this swan song was quite "perfect".

The Super Heavy Booster B15.2, serving as the first - stage rocket, had been launched once during the eighth flight of the Starship and was later retrieved by the "chopsticks" of the launch tower. Today was its second launch mission, and 24 out of the 33 engines on the rocket body were reused.

A close - up photo of B15.2 with obvious "blue - tinted" marks from previous launches | Starbase Surfer

A real - shot photo of the engine part of B15.2, where 24 engines had flown before | SpaceX

However, today it didn't return to the launch tower. Instead, after completing the launch mission, it made a smooth soft - landing on the sea. This wasn't giving up on retrieval but an important test: to practice a brand - new and more aggressive landing procedure for the heavier third - generation Starship.

In this new procedure, in the final stage of landing, 13 engines are first ignited for violent deceleration, then 8 engines are quickly shut down leaving 5 running (this is a step that didn't exist before), and finally, the rocket touches the water gently with the 3 central engines.

The ignition stage of the 5 engines inserted in the middle is like a "safety insurance": in case the 3 central engines fail to ignite, the nearby engines can delay shutdown to ensure the rocket has enough power to complete the landing.

Doing this also has an advantage. It can greatly shorten the time of reverse - thrust deceleration (from the previous 18 - 20 seconds to 15 seconds), thus saving precious fuel and reserving performance for carrying heavier loads in the future.

The Starship S38, as the second - stage rocket, also performed quite stably. It successfully entered the "predetermined orbit", released 8 simulated Starlink satellites, and completed the in - orbit ignition test for the third time successfully.

More importantly, it was its performance when re - entering the atmosphere. In the final stage, it carried out a "dynamic banking maneuver", which is essentially a controllable sideways flight to test the spacecraft's maneuverability at subsonic speeds. It was equivalent to a retrieval drill, practicing how to accurately aim at the virtual "chopsticks" above the sea.

Finally, this spacecraft, which had some heat shields deliberately removed to test the strength of the rocket body, successfully withstood the test of the atmosphere and made a soft - landing in the predetermined sea area.

It can be said that the second - generation Starship ended its era with an almost perfect performance.

However, a glaring fact made this "perfection" seem a bit embarrassing.

Although the process was smooth and the scene was spectacular, this "strongest rocket in history" still failed to truly enter orbit. It was still a "trans - atmospheric sub - orbital" flight.

You can imagine it as throwing a stone as high as possible. It flew very high and even entered space for a while, but it would eventually fall back to the ground and couldn't enter a stable orbit around the Earth like an artificial satellite.

A flowchart of the last flight of the second - generation Starship, still failing to truly enter orbit | SpaceX

Throughout 2025, the second - generation Starship flew five times. The first three flights ended badly, but the last two were relatively successful. However, from beginning to end, it failed to send even one gram of payload into orbit.

This raises an interesting question: seeing that SpaceX finally understood the characteristics of the second - generation Starship, why was it so hasty to "retire" it? What kind of urgency and helplessness were hidden behind this?

Starlink Can't Wait for the Starship Anymore

The answer lies not in the sky but on the ground.

SpaceX is in such a hurry because its other star project, Starlink, is almost "crushed" by its own success.

Starlink, simply put, is a vast space network composed of thousands of satellites. Its goal is to provide high - speed and stable internet access services to any corner of the Earth, whether it's a remote mountain area, the vast ocean, or a signal - blind area.

Starlink is currently the most successful space - based internet satellite constellation | Starlink

You can think of Starlink as a "space information highway" laid for global users. Initially, there were few "vehicles" on the road, and the internet speed was extremely fast. But now, the situation has changed.

As of 2025, Starlink has accumulated over 7 million terminal users in more than 150 countries. The user base of the still - in - its - infancy direct - to - phone service is also rapidly approaching 7 million. With the two - way increase, this "information highway" is becoming overburdened, and "network congestion" has occurred in some areas, with users starting to complain about lag.

Starlink announced that it has over 7 million terminal users and 7 million direct - to - phone users | Screenshot from Starlink's official X account

The only solution is to widen the "road" and add "lanes" - in other words, more Starlink satellites must be launched.

Currently, SpaceX's veteran rocket, the Falcon 9, has been responsible for launching Starlink satellites into space. It's like an extremely reliable pickup truck, working non - stop throughout the year.

The annual launch plan of the Falcon 9 is as high as 144 times, with 80% of the missions allocated to Starlink. Each time, it can carry a maximum of 28 V2 - version Starlink satellites. Even so, the total annual launch mass of Starlink satellites is only 2760 tons, still unable to meet the rapidly growing user demand.

The view from the fairing of the Falcon 9 rocket launching Starlink satellites | SpaceX

But this isn't the most urgent problem. If meeting the needs of existing users is "maintaining business", then the big bet SpaceX just made is staking the future of Starlink.

In September this year, SpaceX spent a whopping $17 billion (about 121.1 billion RMB) to buy 50 MHz of bandwidth in the crucial 2 GHz frequency band (S - band) from EchoStar.

If Starlink is compared to an "information highway", this frequency band is like a legally - protected "exclusive VIP lane" bought specifically for the direct - to - phone service.

Previously, in the direct - to - phone service, SpaceX had to rely on the "lanes" of its partner T - Mobile. Now, they have their own right - of - way.

SpaceX previously cooperated with T - Mobile to launch the Starlink direct - to - phone service | T - Mobile

The problem is, the right - of - way is bought, but the lanes are empty. To activate this VIP lane, a new generation of Starlink satellites equipped with a new communication module must be launched.

According to the plan, the D2D test of this service will start at the latest by the end of 2026. $17 billion has been spent, and the countdown has begun. SpaceX must lay out the new satellite network before then.

The new - generation Starlink satellites are equipped with a direct - to - phone module and inter - satellite laser links, and their communication capacity has been increased to 1 Tbps. However, the satellites themselves have become larger and heavier, each weighing over 1 ton.

Facing such satellites, the Falcon 9 is no longer up to the task.

Therefore, all of Starlink's hopes are pinned on the Starship, because only the Starship can deploy the new - version satellites.

An artistic rendering of the Starship deploying Starlink satellites in orbit like dealing cards | SpaceX

The subsequent Starlink needs to deploy 60 new - version satellites weighing over 1 ton at a time. However, the actual low - Earth - orbit carrying capacity of the currently successfully - tested second - generation Starship is only 35 tons.

It's not a heavy - duty truck but more like a large "minivan".

Now, can you understand SpaceX's sense of urgency? When you urgently need a heavy - duty truck to transport 60 tons of goods and a $17 - billion bet is waiting to be fulfilled, the "minivan" you spent so much effort building, no matter how stable it runs, can't solve the fundamental problem.

Therefore, the second - generation Starship must "retire" because it can't complete the most urgent and crucial tasks for SpaceX now. Its retirement is to make way for the "big guy" that can truly carry hope, the third - generation Starship.

In the tenth test flight, the simulated Starlink satellites collided significantly when leaving the cabin. This problem was perfectly solved in today's eleventh test flight | SpaceX

The Third - Generation Starship: A Complete Makeover

With the second - generation Starship's swan song, SpaceX has devoted all its energy to the research, development, and manufacturing of the next - generation, the third - generation Starship.

This is not just a minor fix but a comprehensive innovation from the inside out, from the first stage to the second stage. The upgrade span is even greater than the improvement from the first - generation to the second - generation.

The core goal is only one: on the premise of full reusability, increase the low - Earth - orbit carrying capacity of the Starship from 35 tons of the second - generation to a brand - new level of over 100 tons.

A comparison of the evolution iteration diagrams of the old and new versions of the Starship | SpaceX

To achieve such a huge leap, the third - generation Starship has undergone many drastic design changes.

A Complete Upgrade of the Power Core

The third - generation Starship will be fully equipped with a new "Raptor 3" engine.

The Raptor engine is the power heart custom - made by SpaceX for the Starship. On the second - generation Starship tested this year, the Raptor engine had evolved to the second generation. Although its parameters were impressive, it experienced in - flight shutdown failures during several launches, and its reliability was far from mature.