Written by: Little Cake, Trend Research

On June 4, SpaceX officially launched its IPO roadshow. A 62-page PPT, 5.556 million shares, each priced at $135, raising $75 billion, with a target valuation of $1.75 trillion. If everything proceeds as planned, pricing will occur on June 11, and trading on Nasdaq will begin on June 12, with the code SPCX.

This will be the largest IPO in the history of human capital markets, surpassing Saudi Aramco, Alibaba, and everything else.

Goldman Sachs, Morgan Stanley, Bank of America, Citigroup, and JPMorgan Chase, five major investment banks, are co-managing the underwriting, with 21 institutions participating in the distribution. Musk himself has a lock-up period of 366 days, while other insiders will see their shares gradually released starting from Q2 2026's financial report, and Fidelity is opening subscriptions to retail investors with accounts holding over $2,000.

The internal code name for the roadshow PPT is Project Apex, which, based on the content, is aptly named.

Three Pillars: Space, Connectivity, AI

SpaceX defines itself in the roadshow as "the only company building infrastructure in space, connectivity, and AI simultaneously." This is not just marketing language; based on financial data, the growth curve, profitability characteristics, and capital demand of the three business segments are completely different, forming a highly complex investment target.

Space: Ground Base

By 2025, SpaceX will complete 165 Falcon series launches, using only 8 newly manufactured boosters. Rocket reuse technology has moved from the experimental stage to industrial production, directly driving down launch costs from the historical average of $18,500/kg to $2,700/kg for Falcon 9 and $1,400/kg for Falcon Heavy, with the goal of Starship V3 to reduce it by over 99% further.

Over 80% of the global orbital mass is transported by SpaceX. In 2023, this figure is 65%, compared to 45% in 2021. This level of market concentration is extremely rare in any infrastructure industry.

However, the financial performance of the space business itself is not impressive. The revenue for 2025 is projected to be $4.1 billion (only from external customers, excluding internal satellite launches), with a year-on-year growth of only 8%. More critically, R&D expenditures for Starship will reach $3 billion in 2025, directly dragging the operating profit of the space segment into a loss of $657 million. Adjusted EBITDA will decline from $1.2 billion in 2024 to $700 million in 2025.

The value of the space business lies not in how much it earns itself, but in its ability to provide the other two segments with deployment capabilities at costs far lower than competitors. Every Starlink satellite launch, every future orbital AI satellite deployment, is built on the cost curves of Falcon and Starship.

Starlink: Cash Cow

Starlink is the true valuation anchor of SpaceX.

In 2023, there are 2.3 million users, projected to grow to 4.4 million in 2024, 8.9 million in 2025, and by Q1 2026 it has reached 10.3 million. It covers 164 countries and regions, with a median download speed of 225Mbps, a median latency of about 25 milliseconds, and an average online rate of 99.9%. Among all maneuverable satellites globally, Starlink accounts for about 75%.

The financial data is even more direct: Starlink is projected to generate $11.4 billion in revenue in 2025, a year-on-year growth of 50%, with adjusted EBITDA reaching $7.2 billion and operating profit of $4.4 billion. This is the only segment of SpaceX that is consistently profitable, and its profit margins are expanding.

The roadshow PPT reveals a key technological upgrade: V3 satellites. Each V3 satellite has a bandwidth of 1024Gbps, more than 10 times that of the current V2 satellites. Using Starship to launch V3 satellites can carry 60 satellites at a time, increasing network capacity by 61,000Gbps in a single launch, which is more than 20 times that of the current Falcon 9 launching V2 satellites.

The V3 satellite is planned to be deployed on Starship in the second half of 2026. If Starship achieves operational-level reusability on schedule, the bandwidth expansion speed of Starlink will leap a magnitude higher, further widening the gap with all competitors.

Starlink Mobile (satellite direct-to-mobile) is also worth noting. Currently, about 650 first-generation mobile satellites have been deployed, covering approximately 1.9 billion people, with partnerships established with about 30 mobile operators, including the announced onboard connection agreement with American Airlines in 2025. The second-generation mobile satellites are planned to be deployed on Starship in 2027, at which point 5G-level speeds and voice services will be offered. SpaceX has also signed an agreement in 2025 to acquire EchoStar's US and global mobile satellite spectrum license for $65MHz, expected to be completed by November 2027.

The roadshow PPT provides a Total Addressable Market (TAM) for connectivity businesses: $1.6 trillion (broadband $870B + mobile $740B). According to the current growth trajectory, the penetration rate of this TAM is still very low.

AI: Money Burning Black Hole or Trillion Dollar Bet?

In February 2026, SpaceX completed the acquisition of xAI through an all-stock transaction, with a post-merger valuation of $1.25 trillion. This transaction is a key variable in understanding SpaceX's $1.75 trillion IPO valuation and is also the major source of controversy.

The merged AI business consists of three components:

The first is computing power infrastructure. Colossus I and Colossus II combined have a computing power capacity of 1GW, billed as the world's largest coherent supercomputer, and the first GW-level cluster with the deployment of GB200 and GB300, complemented by a GW-level Tesla Megapack battery storage system.

The second is the Grok large model. The roadshow PPT states it has reached cutting-edge levels on benchmarks like scientific reasoning (GPQA Diamond), "faster than any other leading model provider." The current version Grok 4.3 was released in May 2026, and SpaceX has also signed a cooperation agreement with Cursor, holding an option to acquire Cursor at an implied valuation of $60 billion.

The third is the X platform. It has about 550 million monthly active users (including Grok and X users), with approximately 350 million posts per day. 117 million monthly active users have used Grok's AI features. X is launching a new advertising platform, X Ads Manager, and plans to evolve into an "all-in-one app" that integrates information, communication, media, payments, and banking.

The monetization of the AI business is currently divided into three lines: consumer side (X Premium subscription + advertising), enterprise side (Grok Enterprise/API + Cursor cooperation), and computing power sales (calculating capacity agreements with cloud service providers at a monthly fee of $1.25 billion, valid until May 2029).

However, the financial reality is harsh. The AI segment is projected to generate $3.2 billion in revenue in 2025, with most revenue coming from X's advertising and subscription income. Adjusted EBITDA will be negative $1.2 billion, operating losses of $6.4 billion, consuming 61% of the company's capital expenditures. Morningstar forecasts that xAI will burn through $10 billion in 2026.

The roadshow PPT gives a near-term TAM for the AI business of $3.8 trillion (infrastructure $760B + consumer subscriptions $600B + digital advertising $2.4 trillion). Including the "larger opportunities unlocked by AI," the total TAM surges to $26.5 trillion.

The Most Valuable Page: Sending GPUs to Space

Pages 35-36 of the roadshow PPT are the two pages with the highest information density in the entire document, and also represent the greatest differentiation SpaceX brings to its investment narrative.

The core logic is that the US ground power supply cannot keep up with the growth in AI computing power demand. In 2025, data center power demand is 62GW, with only 49GW of supply, resulting in a shortfall of 13GW. Between 2008 and 2023, power production in the US has seen almost zero growth, while China has seen a growth rate of about 6% during the same period. Building data centers on the ground faces a series of bottlenecks, including grid approvals, land planning, and community opposition.

SpaceX's solution: move AI computing to space.

The design logic for orbital AI computing satellites is based on the technology platform of Starlink V3 satellites. The roadshow PPT demonstrates a clear evolutionary path: Starlink V3 retains core components such as inter-satellite laser communication links, flight control computers, and attitude control while removing backward antennas, large batteries, and modems, and adding AI computing chips, more solar panels, and larger heat sinks.

SpaceX claims that orbital AI computing has three structural advantages:

First, solar power is unlimited, clean, and cheaper, with allocation achieved through the Starlink network, without the bottlenecks of ground grid approvals. In geosynchronous orbit, satellites are exposed to sunlight for over 99% of the time, supporting continuous AI training tasks.

Second, using radiative cooling is cheaper than liquid cooling or air cooling systems. Data is efficiently routed between the orbital compute cluster and ground users through the existing Starlink network.

Third, the deployment of next-generation chips is faster. Each generation of GPU sees a leap in token efficiency, and through Starship's rapid payload launch cycle, upgrades can occur faster than in ground data centers.

SpaceX estimates that launching one million tons of satellites per year generates 100kW of computing power per ton, increasing AI computing capacity by 100GW per year, with almost no ongoing maintenance costs.

On January 30, 2026, SpaceX submitted an application to the FCC to deploy up to 1 million orbital data center satellites, which was accepted for review by the FCC on February 2. This is the largest data center construction proposal in human history. In the second half of 2026, pilot tests of on-orbit computing nodes will begin using Starlink V3 hardware, with official deployment of AI computing satellites starting in 2028.

The shock value of this narrative lies in re-defining SpaceX from "rocket company + satellite internet company" to "global AI infrastructure provider."

A phrase frequently emphasized by SpaceX in the roadshow is: "Only we can do this."

The confidence behind this statement comes from a vertically integrated chain that other companies cannot replicate: self-developed rockets (reducing launch costs) → self-developed satellites (reducing manufacturing costs) → self-built inter-satellite communication network (reducing data transmission costs) → self-owned AI models (directly consuming computing power) → self-owned end-user platform (X, with 550 million MAU). From chips to space, from space to the terminal, the entire chain is owned.

Google is also making similar attempts. The Project Suncatcher plan announced in November 2025 aims to launch two prototype satellites in early 2027 to verify the feasibility of AI loads operating in-orbit in collaboration with Planet. However, Google needs to rely on SpaceX for launching and lacks the capability to build its own satellite network.

However, the external response to this narrative remains calm.

Calculations by Varda Space Industries show that the current cost per watt for orbital computing is about three times that of ground systems. Musk claims cost parity can be achieved within 2-3 years, but independent analysts generally believe it may only be possible in the 2030s. Cosmic radiation interference with chip computing, heat dissipation engineering in a vacuum environment, and delays between orbital satellites and the ground are all unresolved engineering challenges. The head of Amazon AWS has publicly stated that orbital data centers are "far from being practical."

But even if we look at this story from a discounted perspective, the structural advantages SpaceX holds still stand: If any company wants to do orbital computing, they will ultimately need to purchase launch services from SpaceX. Regardless of whether the timeline for orbital data centers is 2028 or 2035, SpaceX remains the essential pathway.

Pie in the Sky: Developing the Moon and Mars

Pages 43 and 44 of the roadshow have no revenue projections, no timelines, just six phrases, each followed by a scene that sounds like science fiction: lunar economy, Martian energy production and manufacturing, point-to-point Earth travel, in-orbit manufacturing, Mars crewed and cargo transport, asteroid mining...

The PPT title reads: "We are in the best position to create trillion-dollar new markets."

On the lunar economy page, SpaceX provides slightly more details, with three lines advancing simultaneously:

First, in coordination with NASA's Artemis program, return to the Moon by the end of the 2020s using Starship to transport personnel and establish a sustainable lunar base to validate all systems necessary for human long-term survival beyond Earth; Second, build AI satellite factories on the Moon to launch satellites into orbit using solar power and lunar mass drivers; Third, through this manufacturing-launch chain, push global AI computing capacity from GW level to TW level.

Among these three lines, only the first has external endorsement. The NASA Artemis contract is a real commercial order, with SpaceX being the only selected contractor for the human landing system, making sending people to the Moon by the end of the 2020s plausible based on current technological progress.

The second and third lines are still in the engineering concept stage. The list of problems that lunar factories need to address is very long: lunar dust erosion of manufacturing equipment, precision assembly in low-gravity environments, and engineering validation of mass drivers, each of which could take decades to solve.

As for point-to-point Earth travel (using Starship for intercontinental flights from New York to Shanghai in 30 minutes) and asteroid mining, SpaceX itself hasn't provided a timeline.

But the existence of these two pages of the PPT answers a question that investors must face: What exactly does the $1.75T valuation buy?

Morningstar's DCF model values SpaceX at $780B, anchored by Starlink's predictable cash flows and stable revenues from space launch operations. The nearly $1 trillion gap from $780B to $1.75T corresponds to a proposition far greater than five years of profit growth: if human civilization is really to expand beyond Earth, SpaceX is the only infrastructure provider.

This proposition does not require investors to believe that lunar factories will definitely be built, nor that Martian colonization will happen within their lifetimes. It only requires investors to believe one thing: if any of these scenarios even partially becomes a reality, only SpaceX can do it.

Exclusivity, rather than certainty, is the pricing logic behind this $1 trillion premium.

Trend Interpretation

In the view of Trend Research, SpaceX, like Tesla, belongs to the category of faith stocks.

As Morningstar analyzed, Starlink alone can support a valuation of over $600B, but the $1 trillion between $780B and $1.75T is a faith premium, pricing ten-year bullish options like orbital AI computing, lunar economy, and Martian colonization. A 94 times revenue multiple has no precedent among trillion-dollar companies.

Moreover, xAI is the biggest risk variable in this IPO, and discussions about it in the roadshow are far from sufficient.

In the $250B all-stock acquisition in February 2026, Musk controlled both sides of the deal, and this related party transaction made SpaceX instantly shoulder the full losses of the AI segment. Before the merger, SpaceX had briefly been profitable by $800 million in 2024, but after the merger, faced a net loss of $4.9 billion in 2025, followed by a single-quarter loss of $4.3 billion in Q1 2026. The annual operating loss of the AI segment is projected to be $6.4 billion, with an expected $10 billion burn in 2026, while Starlink's $4.4 billion operating profit cannot even cover 70% of this hole.

The position of Grok in the competitive landscape of leading models is far from secure, and the recovery of advertising on the X platform is still in its early stages. More concerning is that Musk maintains absolute control through dual-class voting shares, leaving public shareholders with almost no checks on future related transactions and capital distribution. Morningstar openly states that xAI represents a "substantial threat to value destruction."

Finally, the short-term trading logic of SpaceX may be in complete contradiction to long-term investment logic.

The extremely low free float of 3%, expectations of rapid inclusion in Nasdaq 100 (as early as July), consortium underwriting support from 21 investment banks, and market enthusiasm for AI infrastructure may create an initial demand-supply imbalance that drives the stock price far beyond fundamental support levels.

However, SpaceX's lock-up structure is very unique: insiders can begin selling 20% in batches after the Q2 financial report, with the first full unlock coming in December 2026, and Musk himself is subject to a lock-up period ending 366 days later (June 2027). Coupled with the AI segment's loss figures being progressively revealed in quarterly reports, a significant sell-off window may form from late 2026 to mid-2027.

Overall, the 62-page roadshow PPT depicts an empire of full-stack infrastructure from Earth to space, and from rockets to AI. SpaceX's launch capabilities and the growth curve of Starlink have already proven the execution power of Musk's team. The question is, where are the boundaries of this execution capability? Is it within the atmosphere, or beyond it?

The answer to this question will determine whether the $1.75T valuation is vision or hubris.

Disclaimer: This article represents only the analytical views of Trend Research and does not constitute any investment advice. SpaceX has not officially gone public, and the financial data in the prospectus is preliminary, unaudited, and may be revised. Investors should carefully read the S-1 registration statement and prospectus submitted by SpaceX to the SEC, fully understand the relevant risk factors before making any investment decisions.