SpaceX 为什么拥有如此高的估值天花板？答案藏在马斯克的商业版图里

On June 12, 2026, Eastern Time, SpaceX officially landed on the Nasdaq stock exchange with the ticker symbol SPCX. The company's IPO opening price was set at $135. After opening, the stock price rose steadily and eventually closed at $160.95, surging 19.2% in a single day.

Riding on this epic IPO rally, SpaceX's market cap soared past $2.1 trillion in a single day, setting the record for the largest single IPO in human commercial history (SPCX continued to climb post-IPO, showing the market's boundless imagination for SpaceX's growth potential).

Image: Starship launch photo. Source: www.space.com/

This capital feast directly propelled Elon Musk to the pinnacle of global wealth, making him the first super-rich individual in human history with a personal fortune exceeding $1.1 trillion.

Of course, looking back at Musk's series of operations over the past few years from a longer time horizon, it becomes clear that the SpaceX IPO was actually just a logical step in his vast industrial layout.

Behind this lies a well-planned underlying business logic, where all seemingly scattered actions quietly serve a larger global ecosystem.

Tesla's intelligent manufacturing, xAI's artificial intelligence, Starlink's global network, and Neuralink's cutting-edge technology create layers of foundations for data entry points, manufacturing systems, intelligent computing power, and aerospace technology. They progress step by step, interlocking with each other, and, leveraging capital dividends, continuously integrate, iterate, and empower one another, gradually forming a self-sustaining, ever-evolving complete business closed loop.

In fact, global technological competition today has long moved beyond the competition of single products or point technologies. Future industrial competition will be more about the confrontation of a full-chain ecosystem encompassing computing power, energy, manufacturing, data, and physical execution.

The key to mastering the core voice in the next generation of smart industries lies more in breaking down barriers between different fields and building a complete ecosystem closed loop. And this SpaceX capital feast might just mark the starting point of a new cycle. A deeper technological and industrial competition has truly just begun.

Deconstructing Musk's Empire Ecosystem Map

In fact, over the years, Musk has done many things that lacked validation at the time, or even things people wouldn't dare to imagine. From reusable rockets and global satellite internet to humanoid robots, brain-computer interfaces, and orbital computing power, each project has required enormous investment and long cycles, accompanied by extremely high uncertainty.

If we look at these projects together, we find they are closely interconnected. Musk has been systematically filling all the key capabilities needed for the complete technology system he envisions, centering around artificial intelligence, communication networks, aerospace transportation, intelligent manufacturing, and human-computer interaction.

Currently, I roughly break down this map into four parts:

• xAI and Orbital Computing Power form the intelligent brain;
• Starlink and Starship handle information transmission and physical transportation;
• Tesla and Optimus are responsible for manufacturing and physical execution;
• Neuralink and X connect to neural signals and human societal data, respectively.

These sectors are at different stages of development. Some have already generated stable commercial revenue, some are entering large-scale validation, and others remain in long-term technological exploration.

However, together they form Musk's highly imaginative industrial moat, also extending SpaceX's value boundaries towards communication, computing power, manufacturing, and future space infrastructure.

Image: Musk's Empire Ecosystem Map  Source: www.theinformation.com

The Brain: xAI + Orbital Computing Power

xAI is Musk's artificial intelligence company, best known for the product Grok. However, xAI's role is far more significant than just a chatbot. It also controls large models, supercomputing clusters, and AI infrastructure, acting as the intelligent and computing power hub within Musk's entire technology system.

In February 2026, SpaceX fully acquired xAI, which was valued at $250 billion, further integrating AI with its long-standing aerospace technology and Starlink satellite network.

Since both companies are under Musk's control, many at the time saw this acquisition as a financial arrangement prior to the IPO, a "left-hand-to-right-hand" maneuver to pave the way for SpaceX's listing.

But from a longer-term perspective, this acquisition was primarily aimed at further strengthening the AI and computing power capabilities within the SpaceX system. Upon completion of the integration, SpaceX simultaneously covers space transportation, satellite communications, artificial intelligence, and computing power infrastructure, forming a technology matrix spanning aerospace and AI.

Therefore, we cannot fully view xAI in the same way we understand OpenAI or Anthropic. Grok is just xAI's front-end product for the public; its deeper value lies in providing models, computing power, and intelligent decision-making capabilities for Musk's aerospace, robotics, intelligent manufacturing, and future orbital facilities.

The heavy and unique computing power system behind xAI is also one of its most fundamental differences from ordinary AI companies.

From the perspective of conventional computing power clusters, according to xAI's official disclosures, its Colossus computing cluster has deployed 200,000 H100 GPUs. The entire cluster was initially built in just 122 days, and its scale was doubled in a subsequent 92 days, setting a record for extremely fast construction.

Image: Actual photo of the xAI Colossus supercomputing cluster. Source: www.naddod.com

This means xAI has entered the most capital-intensive and asset-heavy global AI computing power competition, building its own intelligent iteration capabilities from the ground up.

Relying on top-tier computing power, xAI can perform billions of uninterrupted virtual simulations for various hardcore physical scenarios, such as rocket combustion parameters, robot motion trajectories, space material degradation, and space station construction. It can sift through a vast number of options to identify the optimal path for implementation, providing precise intelligent support for the physical operations of the entire system.

However, the iterative upgrade of ground-based AI computing systems has already hit a natural physical bottleneck, an inevitable constraint of technological development.

AI supercomputing research data shows that while the performance of cutting-edge AI supercomputers roughly doubles every 9 months, the corresponding hardware costs and electricity demands also double annually.

For top-tier clusters like Colossus, industry estimates suggest a hardware cost of around $7 billion and an operating power consumption of up to 300 MW. They face four major challenges: energy consumption, heat dissipation limitations, land resources, and network latency. This means the iteration ceiling for ground-based data centers is limited; simply stacking more GPUs or expanding server rooms cannot achieve a qualitative breakthrough.

It's like trying to put things into a fixed-size warehouse; no matter what you do, there's a limit to what you can store.

So, the core reason for Musk's layout of orbital computing power is precisely to break free from the shackles of ground-based computing development and shift to space.

Space offers an inexhaustible supply of free solar energy and a naturally low-temperature environment for cooling with low energy loss. Deploying computing power clusters in low Earth orbit can completely break free from the hard constraints of terrestrial resources, providing a continuous core driving force for the ongoing iteration of AI.

So you see, in recent years, Musk has been frantically launching satellites, partly to forge his space computing network and prepare for the subsequent space computing system.

Furthermore, a Reuters report indicates that SpaceX plans to complete an orbital AI computing demonstration as early as the end of 2027 and has already received approval to launch up to 1 million space data center satellites (Musk's satellite launch costs are extremely low, which we'll detail later, meaning only Musk can do this; others essentially cannot).

In March of last year, xAI acquired the social platform X. One purpose of acquiring X was data. X platform daily accumulates vast amounts of real-time human behavioral traces, group preferences, and social dynamics data. Combined with xAI's own accumulated physical scene simulation data, this intelligent system can fully grasp the complete operational logic of both the physical world and human society.

Compared to the static, lagging, sampled datasets commonly purchased by peers, Musk's system generates real-time, authentic, multi-dimensional endogenous data, creating an irreplaceable differentiated iteration advantage.

Neural Logistics Core: Starlink + Starship

Starlink is a low Earth orbit satellite internet system built by SpaceX. It uses a large number of near-Earth orbit satellites to provide broadband internet globally, especially covering remote areas, seas, and airspace that traditional communication networks cannot easily reach. It acts more like a global communication network built by SpaceX in space and is now widely adopted.

For example, during the Russia-Ukraine conflict, after Ukraine's ground communication facilities were damaged, it relied on Starlink's network services to maintain military command, drone operations, and government communications. After Hurricane Helene caused internet outages in parts of the US in 2024, rescue departments deployed numerous Starlink terminals to restore emergency communications.

Starlink has actually achieved high commercial success. In 2025, SpaceX's sales reached $18.67 billion, with Starlink contributing about 60% of the revenue, making it the group's core cash flow source. Currently, Starlink has over 10.3 million global users and approximately 9,600 satellites in orbit. It has grown from an experimental project into a mature, stable core infrastructure.

Of course, Starlink's core value has long transcended ordinary satellite broadband services. In essence, it serves as the global real-time information network for the entire ecosystem within Musk's system.

Unlike the public perception of it as a "substitute for terrestrial networks," Starlink's core advantage lies in complementary empowerment.

Traditional terrestrial fiber optic networks rely on glass media for transmission, suffering from high latency, significant signal loss, and strong geographical limitations, making them unsuitable for the millisecond-level global coordination needs of advanced AI.

However, low Earth orbit satellite networks equipped with inter-satellite laser links can bypass some path limitations of submarine cables in intercontinental long-distance communications, achieving lower latency with shorter transmission paths. They can build a unique network advantage in scenarios such as global coverage without dead spots, connecting remote areas, extreme environment communications, and low-latency intercontinental transmission, ensuring the efficient linkage and precise operation of this system.

With Starlink, the future orbital computing center can maintain low-latency interaction with ground data systems. For example, an AI inference request initiated from the ground can be uploaded to the space computing center via Starlink for computation, and the inference result can be transmitted back to the ground in real-time via Starlink.

Starship is SpaceX's continuously developed next-generation super-heavy launch system, responsible for sending crew, satellites, and large equipment into space. The "chopsticks catching a rocket" we saw earlier was Starship's recovery test. After launch, the first-stage booster automatically flies back to the launch tower, where two giant mechanical arms catch it directly. This minimizes refurbishment time and enables rapid reuse, a recovery system that significantly reduces Starship's launch cost.

Image: The moment Starship's "chopsticks" catch the rocket. Source: san.com

Although Starship is still in the testing phase and hasn't established a stable commercial launch price, Musk has previously stated that the mature single-launch total cost could be reduced to less than $10 million, with the long-term marginal cost potentially approaching $2 million.

What does this mean? SpaceX's current Falcon 9 standard commercial launch price of about $74 million is already considered very low cost. For context, NASA's SLS single mission cost is as high as $2 billion to $4 billion.

Therefore, Starship, with such low costs, will be the world's only scalable, low-cost, repeatedly reusable space transport vehicle, capable of delivering over 100 tons of payload to low Earth orbit. Traditional space launches are prohibitively expensive and occur infrequently, completely unable to support large-scale commercial space deployment. Starship, however, relies on technology reuse, large-scale manufacturing, and high-frequency iteration to drastically reduce the cost of space operations.

Relying on its superior payload capacity and low-cost advantage, Starship can efficiently execute core tasks like batch deployment of orbital computing nodes, large-scale Starlink satellite constellation assembly, space equipment maintenance, and material transfer between space and Earth.

Starlink handles the rapid flow of information, while Starship handles low-cost physical deployment. One handles the virtual, the other the physical; one handles signals, the other objects. Together, they completely open the two-way communication channel between space and Earth, allowing Musk's ecosystem to entirely break free from the competitive limitations of traditional ground-based technology.

The Physical Body Core: Tesla + Optimus

We don't need to introduce Tesla, the electric vehicle company, too much.

In January 2026, Tesla officially announced the permanent discontinuation of its two flagship models, Model S and Model X. In fact, these two models were once Tesla's flagship products and a core business with stable high gross margins. However, their sales declined, industry competition intensified, and they long occupied significant R&D resources, production line capacity, and core talent, while their value in empowering the global smart closed-loop layout continuously weakened.

Image: Fremont factory group photo including the last two Model S / Model X units. Source: cdn.shopify.com

Authoritative media outlet Axios reported that Tesla's core purpose in stopping Model S and Model X production was to free up premium capacity and factory space at its Fremont factory, fully shifting focus to the development and mass production of the Optimus humanoid robot. Similarly, The Guardian explicitly pointed out that the essence of this product line adjustment is Tesla's corporate positioning iteration, i.e., transforming from a traditional electric vehicle company into a "physical AI company."

In fact, a car is essentially an intelligent robot on wheels, while Optimus is a general-purpose robot that walks on two legs. Their underlying logic is completely interoperable, sharing perception algorithms, intelligent decision-making, motion control, supply chain systems, and mass production capabilities. The core reason for discontinuing the flagship models is to concentrate all premium resources on empowering the iteration and deployment of Optimus.

Image: Full-body photo of the Tesla Optimus humanoid robot. Source: tesery.com

It's not a secret that Musk loves humanoid robots, and he has high hopes for Optimus. Optimus itself is far from an ordinary consumer tech product. It is a general-purpose industrial worker suited for the entire industrial chain, capable of handling high-precision, repetitive, high-risk tasks such as aerospace equipment assembly, precision industrial manufacturing, and inspection and maintenance of hazardous equipment. In the future, it could also be stationed in space bases to perform various extreme scenario tasks, addressing the physical execution shortcomings of the system.

On the other hand, the real physical data generated during Optimus's full-range operations—such as motion trajectories, environmental parameters, and equipment malfunctions—will be fed back in real-time to the xAI hub. This provides continuous real-world data support for algorithm model training, hardware optimization, and operational plan upgrades.

So, Tesla's mature global supply chain and mass production system lay the industrial foundation for the commercial deployment of robots, forming a complete self-sustaining cycle of hardware production, scenario application, data feedback, and intelligent iteration, allowing AI's virtual computing power to truly materialize into sustainable physical productivity.

Human-Machine Interface Core: Neuralink + X

The other line is Neuralink + X.

I actually formed an impression of Neuralink early on, also viewing it as a highly tech-oriented or even sci-fi company. Neuralink is a brain-computer interface company founded by Musk. Its core idea is to implant a tiny chip into the human brain, read neural signals via electrodes, and convert these signals into operational instructions that a computer can understand.

Its most practical application is primarily helping patients with paralysis or severe mobility impairments control computers, phones, and robotic arms using just their "thoughts." For example, after the chip is implanted, a patient doesn't need to move their hands or feet; they just need to generate the intention to operate in their mind to move a cursor, type, or control external devices.

Put more simply, Neuralink establishes a direct communication channel between the human brain and a machine. In the short term, it is firstly a medical technology to help patients restore communication and mobility. Its long-term goal is to further improve the efficiency of information interaction between humans, AI, and robots.