In 'SpaceX: As AI Burns Money, Is 'Space Computing Hegemony' the Ultimate Killer Move?', Dolphin Research mentions that space data centers are not only the core element of SpaceX's grand narrative but also the largest 'option value' in its future valuation.

To achieve this vision, SpaceX has formulated a seemingly crazy (insane) 'Space Computing Power Deployment Timeline':

1. 2028: First batch of computing power satellites enter orbit: The first batch of orbital AI satellites, 'AI1,' is expected to commence large-scale commercial networking in 2028. These satellites will have a wingspan of up to 70 meters and an average power consumption of 120kW (peak 150kW), resembling floating power stations in space.

'Starship + Self-Built Chip Factory' dual engines: Between 2028 and 2031, SpaceX will launch two groundbreaking infrastructure projects: First, the ultra-high-frequency heavy-lift launches of Starship (V3 can place 100 tons into orbit per launch, with a long-term goal of producing 10,000 ships and launching 10,000 times annually); second, its Texas-based Terafab chip factory (using a 2nm process, with a long-term goal of producing 1TW of computing power annually, approximately 800GW dedicated to space).

2. After 2030: Transport 1 million tons of computing power into space annually: With the combined support of transport capacity and computing power, SpaceX plans to transport 1 million tons of computing hardware into orbit annually within 4-5 years (i.e., 2030-2031). Calculated at 100kW per ton of hardware, this corresponds to an annual deployment capacity of 100GW of new space computing power, with an ultimate goal of reaching 1TW (1,000GW).

For comparison, the total cumulative AI computing power deployed on Earth by major cloud service providers (CSPs) globally is currently only in the 30-50GW range. This means that SpaceX's annual 'space computing power increment' alone is equivalent to recreating two to three 'global cloud computing totals' in space. If this plan materializes, it will completely break through the energy and land growth ceilings faced by ground-based computing power.

Facing such a disruptive industrial landscape, Dolphin Research will focus on the following two core questions in this study:

1) Is the transition from ground-based operations to 'space computing hegemony' a dazzling interstellar sci-fi or a 'dimensional reduction strike' against traditional tech giants?

2) Facing such a massive and unprecedented commercial closed loop (ecosystem), how should we value SpaceX, this super unicorn?

Here is the main text:

I. The Soul-Searching Question: Can Vacuum Heat Dissipation Work? Almost There

Ground-based AI data centers are already challenging, but space presents even greater difficulties due to the vacuum environment, which lacks convection and relies solely on thermal radiation to emit heat into deep space in the form of infrared radiation. However, under the same temperature difference, the heat dissipation efficiency is only 1% of that of ground-based air convection.

Heat dissipation is the first major technical obstacle for space computing power, with priority even higher than deployment costs and space radiation. In a vacuum, 'how to expel heat' is the physical prerequisite for all computing activities.

Currently, SpaceX faces several major dilemmas:

a. Area constraints: Based on physical laws, radiation power increases with higher temperatures, larger heat dissipation areas, and higher surface emissivity, leading to faster heat dissipation. However, at a high cabinet temperature of 70°C, the radiation heat dissipation limit is only 880W/m². A 1.5MW data center requires 2,100m² of heat dissipation plates (about 1/3 of a football field), far exceeding the volume of a rocket fairing.

b. Heat dissipation arrays as targets for space microparticles? Due to their large area, 1mm micro-debris in space can puncture thin heat dissipation walls when colliding at orbital speeds.

Additionally, low-Earth orbit (LEO) satellites experience extreme light-dark cycles with temperature differences exceeding 250°C (+120°C to -160°C) every 90 minutes. Such violent thermal shocks can easily cause chip packaging to crack or pipeline fatigue leaks. Since space cannot be manually repaired, a single puncture and leak will result in complete heat dissipation failure and the entire satellite becoming inoperative.

c. High costs: Currently, the International Space Station adopts a customized aerospace model, with heat dissipation costs reaching $4.5-6.6 million/kW. Even with commercial cost reductions, the pure heat dissipation hardware cost is still $6 billion/GW, nearly double that of ground-based data centers ($3.3 billion/GW).

d. Transportation cost disparity: Based on the current status of the Falcon 9, transporting this 'heat dissipation dead weight' into orbit costs $23 billion/GW (nearly four times the cost of heat dissipation hardware). Even if Starship transportation costs drop to $200/kg in the future, the total launch cost in 2026 (with a specific power of 80W/kg) will still be $2.5 billion/GW. It is not until 2032 (with thermal control iterations and a specific power of 195W/kg) that costs are expected to drop to $1 billion/GW.

Facing these contradictions, space data centers must find a balance between 'efficiency, weight, and reliability':

a. Trade lifespan for area (increase temperature tolerance): Utilizing the physical characteristic that radiation efficiency is proportional to the fourth power of temperature, chips are allowed to operate at full capacity at 85-100°C. For every 20°C increase in temperature, the heat dissipation area can be compressed by 15%-25%. The trade-off is reduced reliability and accelerated chip depreciation (GPUs and HBMs operating above 85°C for extended periods will accelerate failure modes).

b. Trade power consumption for space (active liquid cooling decoupling): Adopting a 'cold plate → active pump → coolant → external radiation plate' transport path. Although this adds an additional 2%-4% power consumption and the risk of pump failure, it removes the geometric constraint that chips and heat sinks must be 'in close contact.'

c. Trade weight for cost (material downgrading and folding deployment): Abandon expensive aerospace materials and use ordinary 6061-T6 aluminum alloy with good thermal conductivity but higher weight, following the logic of 'cheap dead weight for low manufacturing costs.' During launch, it can be folded like an 'accordion' and then deployed on a large scale after entering orbit.

d. Trade redundancy for risk resistance (independent modular honeycomb piping): Reusing the engineering experience from Starlink, integrated chassis with radiation fins are used, and the liquid cooling piping is designed as an independent modular honeycomb network. In the event of an impact, a leak in a single piping section can be instantly physically isolated, effectively preventing systemic total loss caused by a single point of failure.

From a technical perspective, the active liquid cooling + deployable heat sink approach is theoretically feasible but still in the engineering verification stage and has not undergone large-scale deployment testing.

II. Space Radiation: Will It Penetrate Chips? Not a Big Problem

In semiconductor physics, the core indicator determining whether a transistor will be affected by radiation is the 'critical charge' – the minimum energy required to trigger a transistor state flip (0→1).

As chip processes evolve from 28nm to 3nm and smaller, transistor volumes shrink dramatically, and operating voltages decrease significantly, causing the critical charge to decline exponentially.

High-energy particles in space can easily cause single-event upsets (SEUs, data errors) and single-event latch-ups (SELs, short-circuit burnouts). However, traditional radiation-hardened large-process chips lack sufficient computing power for AI tasks.

SpaceX's solution is to accept localized errors while ensuring system stability:

a. Orbital advantages: Deploying in 500–1000km LEO/SSO orbits leverages the Earth's magnetic field to deflect most high-energy particles, reducing radiation flux at the source.

b. Heterogeneous architecture separation: Using 3nm GPUs for computation (the 'brain') and 65/28nm radiation-hardened FPGAs/MCUs for monitoring (the 'cerebellum') to detect abnormal currents in real-time and cut off/restart GPUs within milliseconds, preventing burnout risks.

c. Spot gradient shielding: Abandoning full-cabinet heavy metal wrapping and instead applying an extremely thin coating of 'low-Z polymer + high-Z tantalum/tungsten' only above core GPU and power management chips to suppress secondary radiation while balancing thermal conductivity and lightweight design.

d. AI's natural tolerance + hierarchical fault tolerance: LLMs are probabilistic models, and single-point SEUs are acceptable in most inference scenarios. HBMs are equipped with ECC for automatic error correction, and core control nodes deploy triple modular redundancy (TMR) for majority voting, completely filtering out single-point hard errors.

Google's research, through 67MeV proton beam experiments simulating extreme LEO radiation, has shattered the traditional belief that 'space must use expensive specialized chips':

HBM memory (3x tolerance, imperceptible error correction): It absorbed 2 krad (nearly three times the expected 5-year dose for ultra-low-orbit satellites) before individual errors occurred, all of which were automatically repaired by ECC with no business impact.

Core computing chips (20x tolerance, zero physical damage): They withstood 15 krad (20 times the expected dose) without any permanent damage, maintaining stable AI training and inference tasks throughout.

This testing empirically validates that the technical route of 'advanced process (COTS chips) + software fault tolerance (ECC/watchdog reset)' can withstand extreme tests.

III. Latency: A Problem!

Internally, space data centers remain standard 'NVIDIA data centers,' but their external connectivity relies on a vast wireless network woven from 'space lasers (inter-satellite high-speed links)' and 'microwave/hybrid space-ground laser links (space-to-ground backhaul)':

a. Internal interconnectivity (within satellites vs. ground cabinets): Identical

GPUs on the same motherboard still use NVLink/NVSwitch for interconnectivity, and different compute nodes are still connected via traditional Ethernet or InfiniBand to form local area networks.

b. Node interconnectivity (between satellites vs. between ground data centers): From 'tangible' to 'intangible'

On the ground: Data centers or cabinets must rely on physical fiber-optic cables buried underground or suspended in the air for connectivity.

In space: Completely wireless. Different satellites use optical inter-satellite links (ISLs), which employ invisible laser beams for ultra-high-speed data transmission.

c. Backbone backhaul (between space and ground vs. ground backbone networks): A dilemma of stability versus speed

This is the most significant difference between space and ground networks. Ground data centers directly access extremely stable, ultra-high-bandwidth ground fiber-optic backbone networks. In contrast, space computing power must transmit data to Earth users across the thick atmosphere, facing physical-level compromises:

Stability-first (Ka-band microwaves): The current primary solution. Using electromagnetic waves for data transmission is slower (about 17Gbps) but 'rugged and reliable,' unaffected by cloudy or rainy weather, ensuring 24/7 connectivity.

Speed-first (optical ground laser links): A future upgrade option. Bandwidth is hundreds of times higher, suitable for transmitting massive AI data. However, it is extremely 'delicate,' prone to disconnection in cloudy or foggy conditions, requiring the construction of a vast number of backup ground stations globally, 'dependent on the weather.'

Currently, space data centers face several issues:

a. Data latency: A low-Earth orbit (LEO) computing satellite orbits the Earth 15 times a day, spending only 5-7 minutes over a specific ground station each time. Connection quality is good only when the satellite is directly above the nearest ground station, which lasts only 5-7 minutes daily.

Once it moves away, data must be relayed multiple times ('Chinese whispers') among satellites, causing one-way latency to surge to 30-80ms (ground fiber optics: <1ms).

b. Space-to-ground backhaul: The problem worsens when switching to optical ground links instead of RF links. Space-to-ground laser links are highly susceptible to interference from clouds and rain, leading to disconnections. Building a large number of ground stations globally is costly, and communication between dispersed ground stations and end-users becomes another source of increased latency.

For SpaceX, feasible solutions include:

b. Promoting 'integrated sensing and computing' edge computing: Enabling satellites to process images and immediately offload them to nearby computing satellites, where AI interprets the results in orbit within 1 second, compressing 10GB of raw images into a few KB of conclusion 'text messages' (e.g., 'abnormal target detected at coordinates X') for transmission back to Earth. This reduces downstream data volume by over 90%.

With significantly reduced data volume, even if laser links are obstructed, they can switch to weather-resistant microwave backup links (Ka/V-band) for all-weather, second-level responses. However, this approach affects AI's multi-round, multi-modal interaction scenarios.

Communication latency stems from physical limitations of light speed and orbital mechanics and cannot be eliminated through technical means. Therefore, space data centers must abandon millisecond-level real-time scenarios (autonomous driving, high-frequency trading) and precisely position space computing power for high-latency-tolerant asynchronous computing: AI training (day/week cycles), meteorological and climate simulations (tolerant of a few seconds' delay), and on-site space computing (debris collision warnings, astrophysical modeling, etc.).

Operations and maintenance: The inability to conveniently intervene manually in orbit is the core operational challenge for space data centers. At present, redundancy design (e.g., preset (preset) 20% GPU overconfiguration to handle irreparable permanent hardware failures and radiation-induced computing availability degradation – e.g., 95% computing availability) and software-level fault tolerance mechanisms (ECC error correction, watchdog resets, etc.) are used to replace on-site repairs, increasing overall deployment and operational costs.

Space robot maintenance is still in the experimental stage. It is expected that after 2032, as in-orbit robotics technology matures, a certain degree of in-orbit repairs and component replacements will become possible, extending the service life of space data centers.

IV. Cost Economic Viability: Is It Reliable?

The above analysis primarily focuses on technical feasibility. Now, let's examine economic viability. Compared to ground-based computing centers, space computing centers primarily leverage abundant energy resources.

However, this energy is not entirely free. Sunlight duration varies significantly across different orbits, directly affecting power supply capacity and energy storage costs:

Low-Earth Orbit (LEO): Orbiting the Earth about 15 times a day, LEO satellites receive sunlight only 60% of the time, resulting in low average effective solar irradiance levels. Frequent entry into Earth's shadow zones requires large-capacity energy storage batteries, significantly increasing system complexity and hardware costs.

Sun-synchronous dawn-dusk orbit (SSO): The preferred orbit for space data centers. Orbiting along the Earth's terminator in a retrograde manner, SSO satellites remain continuously sun-facing throughout most of the year, with a maximum shadow period of only 35 minutes daily. The required battery energy storage capacity is much lower than that of LEO. However, this orbit is a scarce resource with much smaller available capacity than ordinary near-Earth orbits.

The energy for space data centers also follows a CAPEX-replacing-OPEX model: there are no continuous electricity bills, as all energy costs are embodied in the upfront investment for solar arrays and energy storage systems.

The essence of space computing power is not merely arbitraging the "scarce total electricity supply," but rather using high fixed costs—launch, on-orbit system manufacturing, and reliability costs—to hedge against the full range of capacity expansion bottlenecks faced by ground-based data centers. These include not only direct cost increases in power supply but also multiple non-energy constraints such as grid interconnection queues, land and environmental permits, industrial material production capacity, and construction labor.

From a supply hierarchy perspective, ground-based computing power faces four progressive layers of buffering in power supply, each corresponding to different costs and scalability challenges:

Only when these four supply layers are gradually exhausted and the comprehensive costs of ground-based computing power continue to rise do space data centers begin to demonstrate their economic value. Prior to this, there remains significant untapped cost reduction potential on the ground.

To some extent, the core question of its viability hinges on whether the electricity-related costs of operating ground-based computing centers can outperform the additional (non-energy-related) costs required for space-based computing over the operational lifecycle.

Under this framework, whether space computing power represents a "dispensible backup" or an "essential necessity" is expected to follow two distinct evolutionary paths:

1) Stabilizing power supply-demand balance

Cost dynamics: From "prohibitively expensive" to "long-term parity"

Initial disadvantage (2026): The total cost of ownership (TCO) for space data centers exceeds ground-based counterparts by over 4x. High costs stem from custom radiation-hardened/thermal control hardware, shortened chip lifespans due to radiation/thermal stress (5 years vs. 15 years), radiation-induced chip availability degradation (95%), and extreme system redundancy requirements due to non-repairability (requiring 20% GPU redundancy).

Long-term parity (2040): As engineering challenges in thermal management and space radiation are overcome, combined with Starship-driven launch cost reductions, the levelized cost of computing (LCOC) between space and ground will reach parity by ~2040 (in fact, by the early 2030s, space costs will be only 30% higher than ground-based, approaching scalability thresholds).

b. Supply-demand evolution: Ground power remains sufficient, making space "optional rather than essential"

Under baseline assumptions, ground-based power capacity will steadily expand from 89 GW in 2026 to 338 GW by 2030.

2) Severe bottlenecks in ground power expansion (power shortages)

a. Cost divergence: Parity achieved 6 years earlier

Surging ground costs: Driven by grid approval delays and shortages in gas turbine/transformer production capacity, ground data center CAPEX spikes from $34.6M/MW (baseline) to $53.4M/MW.

Plummeting space costs: Starship reduces launch costs to $80/kg with economies of scale, cutting space data center CAPEX to $11M/MW.

This "cost scissors difference " (cost divergence) effect brings space LCOC to parity with ground-based by ~2034 (6 years ahead of baseline), after which space cost advantages continue widening—by 2039, space LCOC will be nearly 20% lower than ground-based, establishing clear cost competitiveness.

b. Supply-demand evolution: Historical inflection point—computing explosion triggers "space spillover effect"

Chip capacity surge: Terafab adds ~1M wafers/month capacity by 2040, significantly raising deployable computing power ceilings.

Ground power capacity peaks in 2028 and stagnates thereafter, reaching only 576GW by 2035 (vs. 1,150GW in baseline) and ~2,400GW by 2050 (vs. ~7,500GW in baseline). Computing expansion constraints shift definitively from chip capacity to ground power infrastructure.

Space spillover effect activates: Despite cost parity in 2034, true demand spillover occurs in 2037 when total chip capacity breaches ground power capacity limits, forcing massive computing deficits to shift into space.

As demand continues growing, space computing scales rapidly: ~200GW in orbit by 2038, surging to ~4,800GW by 2050 (accounting for nearly 73% of annual chip capacity).

At this point, space data centers cease being supplementary options and become the core—nearly exclusive—viable solution for large-scale AI computing deployment.

V. How to Value SpaceX, This Behemoth?

① Rocket launch business: In "From Pipe Dreams to Golden Fortunes: Is SpaceX Really That 'Sci-Fi'?", Dolphin Research argues rocket launch represents near-absolute monopoly, with some comparing it to the East India Company during the Age of Exploration.

For this unique monopolistic asset, our valuation approach:

Assuming full utilization of 1 million tons annual launch capacity, priced at market fair value of $200/kg, Starship's long-term annual revenue could reach $200 billion.

Profitability: After initial R&D and test flight investments, referencing Falcon 9's ~30% EBITDA margin under stable commercial operations, this business could generate ~$60 billion in annual EBITDA long-term. More detailed analyses are published in the Changqiao App [Dynamic-Depth] section under the same article title.

② Starlink business: In "SpaceX: The Celestial Network—Unstoppable?", we noted it essentially leverages space transportation monopoly to create a "space-based version of a telecom carrier".

The core challenge lies in spectrum "spatial reuse". While terrestrial operators can densify urban base stations to reuse spectrum, a single satellite's beam covers tens of thousands of square kilometers, forcing all users in the region to share limited bandwidth.

Thus, Starlink complements rather than replaces traditional terrestrial operators. Its core market will long remain suburban/remote areas and maritime/aerial scenarios where ground networks struggle to cover or face extremely high deployment costs, rather than penetrating dense urban centers.

SpaceX plans to complete deployment of a 42,000-satellite mega-constellation by 2030 (currently ~9,600 satellites in orbit), covering multiple generations from V1, V2 to V3 and V-band satellites.

Among this 42,000-satellite target, next-gen V3 satellites represent the current strategic upgrade focus. Compared to the mainstay V2 Mini, V3 satellites deliver comprehensive improvements in bandwidth, latency, and hardware architecture:

a. Broadband business: According to ARK Invest, global suburban and remote populations total 3.45 billion people (~800 million households). However, even with a 42,000-satellite V3 constellation (1 Tbps per satellite), physical capacity limits cap maximum coverage at 1.63 billion rural residents (~380 million households) and 50 million urban residents (~13 million households).

Based on this, we introduce an extreme assumption for Starlink's long-term expansion:

Assuming continuous capacity expansion eventually covers all 3.45 billion rural residents globally, with "localized pricing"—monthly broadband fees set at 2% of regional per capita GNI (industry affordability benchmark).

Under this ultimate scenario, Starlink's theoretical total addressable market (TAM) for global lower-tier markets could surge to $249.6 billion annually (implying ~$26/month per household ASP).

However, actual revenue conversion will face constraints: not all lower-tier markets have broadband demand, traditional ground base stations will continue expanding into suburbs, and geopolitical barriers in regions like China and Russia will limit market access. Starlink cannot achieve winner-takes-all dominance.

Thus, our revenue model assumes: Under neutral expectations, Starlink captures 30% of the global rural market ($74.9 billion revenue); under optimistic expectations, leveraging scale effects and first-mover advantages, it reaches 50% market share ($124.8 billion revenue).

b. DTC business: For long-term valuation of direct-to-cell services, we take SpaceX's disclosed 8 billion global mobile devices as baseline. Referencing ~$8 global mobile ARPU, this market's long-term TAM reaches ~$740 billion.

However, constrained by physical satellite bandwidth limitations in dense urban areas, Starlink's DTC positioning isn't to replace traditional carriers but to serve as a "global coverage, blind spot elimination" value-added service, forming B2B2C partnerships with global telecom giants.

Under the current 55:45 revenue-sharing model, Starlink cannot capture full mobile ARPU. Its per-device monetization will transform into "wholesale-like" value-added sharing (e.g., Starlink provides underlying network access for T-Mobile US plans, with T-Mobile billing end-users and sharing revenue proportionally).

Leveraging its early-mover device base and partnerships with global leading carriers, we model penetration scenarios:

Neutral expectation: Assuming Starlink captures 10% of global mobile connections (~800 million devices) through value-added services and roaming partnerships. At 55% revenue share, this could generate $40.7 billion in annual revenue.

Optimistic expectation: With deepening global demand for seamless connectivity, assuming device penetration rises to 20% (~1.6 billion devices), DTC revenue could reach $81.4 billion annually, establishing it as SpaceX's explosive second growth engine.

c. Aviation + maritime business

In high-value enterprise scenarios, aviation and maritime broadband represent Starlink's premium-priced, high-margin segments. Estimates:

Aviation market: ~30,000 global commercial aircraft, with extremely high ARPU (~$300,000 annually), implying $9 billion theoretical annual revenue.

Maritime market: ~100,000 active merchant vessels, at ~$34,000 annual ARPU, implying $3.4 billion theoretical annual revenue.

Combined, this niche segment's total addressable market (TAM) reaches $12.4 billion.

Leveraging LEO constellation advantages in "low latency, high bandwidth, global seamless coverage", Starlink is rapidly dismantling traditional GEO satellite providers' barriers.

Assuming long-term stabilization where Starlink captures 80% of addressable aircraft and vessel markets through hardware cost and experience advantages, this segment could stably contribute ~$10 billion in high-margin annual revenue.

Combining penetration projections across these three core businesses, we value Starlink's 2030 long-term prospects using enterprise value multiples (EV/EBIT) and discounted cash flow (DCF) methods (discounted back to 2026 baseline). More detailed analyses are published in the Changqiao App [Dynamic-Depth] section under the same article title:

Neutral expectation: Starlink could reach ~$128 billion total revenue by 2030. Given satellite networks' extremely low marginal costs and scaling scale effects, we assume 45% operating profit margin (OPM), implying $57.6 billion EBIT.

Optimistic expectation: With global penetration exceeding expectations, Starlink could hit $218.6 billion total revenue by 2030. With stronger scale effects, assuming OPM rises to 50%, EBIT could reach $109.3 billion.

③ AI Business: No Unique Value

In 'SpaceX: AI Burns Money Endlessly, Is 'Space Computing Power Hegemony' the Ultimate Killer Move?', Dolphin Research mentioned that SpaceX's AI business comprises the X platform, the Grok model, Colossus ground computing power leasing, and the space data center business:

a. X Platform

Although the X platform is still in the process of reconstruction, it is an indisputable fact that its current advertising revenue has declined from a high of $2.3 billion in 2023 to $1.8 billion in 2025. The X platform is currently trapped in the dilemma of being the 'American version of Weibo': Although it remains the center of public opinion during sudden major events, its daily commercial traffic and user stickiness are being systematically eroded by competitors, and its market share is under continuous pressure.

b. Grok Model

Since SpaceX has not separately disclosed the revenue generated by the Grok model, Dolphin Research has made estimations based on publicly available information:

C-end Revenue Estimation:

1.9 million SuperGrok users: Based on a three-tier distribution (assuming Lite at $9/month accounts for 50%, Standard at $28/month accounts for 45%, and Heavy at $265/month accounts for 5%), the weighted monthly ARPU is approximately $30, with an ARR of approximately $680 million.

4.4 million X Premium users: Their subscription fees are primarily for social functions but also include the benefits of Grok. Assuming the incremental value of the Grok function is $8/month, the ARR is approximately $420 million.

The total C-end ARR is approximately $1.1 billion.

B-end Revenue Estimation:

The Grok Business/Enterprise/API business is still in its infancy, and the gap in model capabilities restricts its penetration in enterprise-level scenarios, resulting in minimal monetization volume. Assuming B-end revenue accounts for approximately 10%, the B-end ARR is approximately $100 million.

The current total ARR of the Grok model is approximately $1.2 billion.

c. Ground Computing Power Leasing Business:

Currently, SpaceX's computing power leasing focuses on an ultra-large-scale single-tenant model for 'a few ultra-large customers' and has signed contracts with three major clients. These three contracts alone have contributed $27.8 billion in ARR to SpaceX's AI business.

As Dolphin Research has also mentioned, SpaceX's computing power leasing is a highly profitable business—deploying computing power with a total investment lower than the industry average while leveraging the advantages of scarce assets and risk clauses to achieve pricing power 3-4 times higher than the industry average, locking in profit margins far exceeding those of peers.

However, the sustainability of this high profitability also faces constraints: ① The 90-day termination clause means that super contracts may disappear at any time; ② After computing power supply and demand balance out after 2027/2028, pricing premiums will be compressed. Therefore, this is more of an 'extremely short-term windfall' during a scarce window period rather than a perpetual business that can be linearly extrapolated.

d. Space Data Center Business:

As mentioned above, SpaceX plans to achieve the ambitious goal of deploying 100 GW of computing power in orbit annually. Referencing the current pricing benchmark of approximately $10 billion/GW in the new cloud computing power leasing market, the total annualized revenue of this business under full-load operation would reach as high as $1 trillion.

Considering that this business, once matured, will possess monopoly and heavy-asset attributes similar to those of a 'space utility,' we assign it a steady-state net profit margin of 20% (corresponding to a net profit of $200 billion) and a relatively conservative 10x PE valuation multiple. Under this benchmark, the terminal market capitalization anchor for the space data center is $2 trillion.

However, due to differences in macro constraints, the timing for this '$2 trillion terminal asset' to truly materialize and deliver profits varies significantly. Combining an industry-weighted average cost of capital (WACC) of approximately 10%, we discount it to the current point in 2026:

Neutral Expectation (Baseline Scenario Extension): Space Computing Power as an 'Option'

If ground power expansion is sufficient to absorb chip production capacity (capacity), the space data center will lack short-term irreplaceable rigid demand (rigid demand) attributes and will serve more as a strategic reserve for ground computing power. In this scenario, we assume that the large-scale deployment node of 100 GW will be postponed to 2045.

Optimistic Expectation (Musk Scenario Realization): Space Computing Power as a 'Must-Have'

If ground power shortages intensify and wafer production capacity (capacity) hits a ceiling, the 'space spillover effect' will be forcibly activated. The space data center will become the core foundation for bear (supporting) the global AI computing power explosion.

In summary, the space data center is not only an engineering marvel but also a super bullish option catalyzed by 'Earth's physical bottlenecks'—the lower the ceiling on the ground, the faster SpaceX will soar toward a trillion-dollar market capitalization.

- END -

// Reprint Authorization

This article is an original piece by Dolphin Research. Reprinting is only allowed with authorization.

// Disclaimer and General Disclosure

This report is for general comprehensive data purposes only, intended for general reading and data reference by users of Dolphin Research and its affiliated institutions. It does not take into account the specific investment objectives, investment product preferences, risk tolerance, financial status, or special needs of any person receiving this report. Investors must consult with independent professional advisors before making investment decisions based on this report. Any person making investment decisions using or referring to the content or information mentioned in this report does so at their own risk. Dolphin Research shall not be liable for any direct or indirect responsibilities or losses that may arise from the use of the data contained in this report. The information and data in this report are based on publicly available materials and are for reference purposes only. Dolphin Research strives to ensure but does not guarantee the reliability, accuracy, and completeness of the relevant information and data.

The information or views mentioned in this report shall not, under any jurisdiction, be considered or construed as an offer to sell securities or an invitation to buy or sell securities, nor shall they constitute advice, inquiries, or recommendations regarding relevant securities or related financial instruments. The information, tools, and materials in this report are not intended for or proposed for distribution to jurisdictions where the distribution, publication, provision, or use of such information, tools, and materials contradicts applicable laws or regulations or results in Dolphin Research and/or its subsidiaries or affiliated companies being subject to any registration or licensing requirements in such jurisdictions, nor to citizens or residents of such jurisdictions.

This report merely reflects the personal views, insights, and analytical methods of the relevant creators and does not represent the stance of Dolphin Research and/or its affiliated institutions.

This report is produced by Dolphin Research, and the copyright is solely owned by Dolphin Research. Without the prior written consent of Dolphin Research, no institution or individual shall (i) make, copy, reproduce, duplicate, forward, or create any form of copies or replicas in any manner, and/or (ii) directly or indirectly redistribute or transfer them to other unauthorized persons. Dolphin Research reserves all relevant rights.