Owning a piece of history.

The Timeline

On April 1, 2026, SpaceX submitted a Form S-1 registration statement to the United States Securities and Exchange Commission (SEC), taking the first formal step toward what will be the largest initial public offering (IPO), ever. SpaceX is targeting a nearly $2 trillion dollar valuation, and plans to raise up to $75 billion, under the ticker symbol $SPCX.

The first public confirmation did not come from a bank or a press release. It came from Eric Berger, the Ars Technica space journalist, who posted on December 10, 2025: "Here's why I think SpaceX is going public soon." Elon Musk replied that afternoon: "As usual, Eric is accurate."

Each milestone in the timeline below carries a specific function: April 1 started the clock with the Securities and Exchange Commission (SEC), the public prospectus, released May 20 exposes the company to retail investors and analysts, the roadshow builds the institutional order book, pricing day fixes the final share price, and June 12 is when SpaceX becomes a publicly traded stock on the Nasdaq Global Select Market.

The exchange choice is itself part of the story. In the Form S-1, SpaceX selected Nasdaq instead of the New York Stock Exchange (NYSE), potentially to utilize Nasdaq's new Fast Entry rule for the Nasdaq 100. The rule lets a qualifying megacap IPO join the index in as little as fifteen trading days after listing, versus the three to twelve months a newly public company normally has to wait.

It is reported that SpaceX is making early inclusion to the index a condition of the listing, and at the nearly $2 trillion dollar target valuation, the company would qualify on day one as one of the six largest U.S. companies by market value, well inside the size threshold that triggers Fast Entry.

The practical impact on passive funds is enormous. The Nasdaq 100 is tracked by exchange-traded funds, mutual funds, and target-date funds holding more than $600 billion in combined assets under management. Fast inclusion triggers automatic, mandatory buying of $SPCX shares to match the new index weight, which creates immediate liquidity, broader institutional ownership, and a steady source of demand that helps absorb any insider selling once the standard 180-day post-listing lockup expires in mid-December. (SpaceX has elected to use a staggered insider share lockup structure that gradually unlocks shares, with all non-Founder-owned shares becoming available 180 days after IPO). The trade-off, which some analysts have flagged, is that this same mechanism forces passive investors into a high-valuation growth stock from day one, before the active market has had time to settle on a fair price.

Recent reporting says BlackRock was discussing a $5-$10 billion anchor commitment in the offering. A check that size, on a single tech IPO, would be one of the largest institutional bets placed on a public listing. When a well-established firm like BlackRock commits to the offering on Day 1, other firms are likely to tail along.

No IPO in the history of capital markets has approached the dollar amount SpaceX is preparing. Saudi Aramco's $29 billion 2019 listing held the record for the largest IPO ever; SpaceX's $75 billion target eclipses Aramco by ~a factor of three. The visualization below stacks the three offerings on a common axis to highlight the differences:

What Is SpaceX?

SpaceX is the American space company Elon Musk founded in 2002. Its full name is Space Exploration Technologies Corporation. The company designs, builds, and launches rockets. It also operates Starlink, the largest commercial satellite constellation ever built. And it is in the process of building Starship, a fully reusable rocket designed to make orbital launch as routine as airline travel. But the company is not stopping. There is more.

The mission of SpaceX is to make humanity multiplanetary by developing fully and rapidly reusable rockets- the pivotal breakthrough needed to substantially reduce the cost of space access.

The launch business sells rides to orbit on the Falcon 9 rocket, and it sells those rides to NASA, the United States Department of War, commercial satellite operators, and a growing list of foreign space agencies.

The Starlink business sells high-speed internet from a fleet of low-orbit satellites to homes, businesses, ships, planes, and increasingly to mobile phones directly.

Starship, the new fully reusable super-heavy rocket, is the system designed to lift the next generation of Starlink satellites, large cargo to the Moon, and eventually crewed missions to Mars. And most recently, a new part of the business- a cutting-edge AI arm called SpaceXAI:

What Is SpaceXAI?

SpaceXAI is the artificial intelligence arm of SpaceX. It builds the language models, computer-vision systems, and inference infrastructure that power the division that ships four flagship products today: MACROHARD, Grokipedia, Grok/ Imagine, and X.

xAI will be dissolved as a separate company, so it will just be SpaceXAI, the AI products from SpaceX

— Elon Musk (@elonmusk) May 6, 2026

Elon's portfolio has undergone quite a bit of restructuring over the past few years: In October 2022, Elon acquired Twitter and renamed it X. In March 2023, he founded xAI as a separate artificial intelligence research company, with the goal of building large language models that compete with OpenAI, Google, and Anthropic. In March 2025, xAI acquired X in an all-stock transaction, putting the social platform inside the artificial intelligence lab. Then in 2026, in a move that surprised most outside observers, SpaceX acquired xAI in another all-stock transaction.

The combined artificial intelligence division will be renamed SpaceXAI. The result is that SpaceX now owns one of the three largest artificial intelligence labs in the world- powered by Colossus- one of the planet's biggest hyperscale AI compute facilities, a social network with hundreds of millions of users, and an full chip fabrication roadmap, all under one corporate umbrella.

MACROHARD

MACROHARD (sometimes referred to as "Digital Optimus") is Elon's jab at "Micro-soft", and the most ambitious bet inside the SpaceXAI product line. MACROHARD is a joint SpaceXAI-Tesla project, coming as part of Tesla’s investment agreement with the former xAI. The premise is that the entire software industry has been built around an old unit of production: human labor.

For sixty years, software has been written by teams of human engineers using human-scale tooling. MACROHARD snaps that paradigm. One human engineer, paired with a fleet of Grok-class artificial intelligence agents, can ship the same utility code that a one-thousand-engineer team used to ship. If the bet works, the cost curve for building software collapses by 3 orders of magnitude or more, and most existing software companies become even more structurally overstaffed overnight.

We are literally painting MACROHARD on the roof of the Colossus II supercompute cluster in Memphis big enough so you can read it from space

— Elon Musk (@elonmusk) September 16, 2025

The @xAI MACROHARD project will be profoundly impactful at an immense scale 😉

Our goal is to create a company that can do anything short of manufacturing physical objects directly, but will be able to do so indirectly, much like Apple has other companies manufacture their…

— Elon Musk (@elonmusk) October 12, 2025

The clearest evidence of how seriously SpaceXAI takes the MACROHARD thesis came with the SpaceX-Cursor deal announced on April 21, 2026. Cursor is the artificial intelligence-powered code editor that became the dominant tool for engineers writing software alongside large language models. The company reached a multibillion-dollar valuation in record time as developer adoption ran ahead of every other AI-coding product on the market.

The structure of the deal is unusual. SpaceX secured a call option to acquire Cursor later in 2026 for $60 billion. If SpaceX decides not to exercise the option, Cursor still receives a $10 billion floor payment in exchange for the joint work. The arrangement gives SpaceX flexibility as it moves through its own Initial Public Offering while giving Cursor a guaranteed $10 billion downside and immediate access to the compute it needed to train its next generation of models.

On the operational side, the partnership is already running. Cursor now has access to the Colossus II supercomputer cluster in Memphis, roughly one million H100-equivalent GPUs, which has dramatically accelerated training of its newest model, Composer 2.5. The combined offering is the practical version of the MACROHARD pitch: one human engineer at the keyboard, a fleet of Grok-class artificial intelligence agents on the other side, and a tooling layer that lets the two work together at speeds traditional engineering teams cannot match.

When MACROHARD becomes commonly known, the entire way knowledge workers operate will be completely abstracted away. Instead of switching between tools, opening and closing apps, and managing endless tabs, the operator will direct intelligent agents that watch the screen, understand the goal, and execute, with Grok serving as the strategic navigator and Tesla's real-time vision system as the fast executor. Microsoft reportedly looked at acquiring Cursor in the weeks before the SpaceX deal was announced but did not move forward, leaving the field clear for SpaceX to lock in the option.

Grok

Grok is the foundational large language model that powers everything else inside the SpaceXAI portfolio. From the start, Elon has framed Grok with a single design goal: a maximally truth-seeking artificial intelligence. The thesis is that the only way to ensure a safe long-term outcome for advanced artificial intelligence is to make truth-seeking the primary objective, even when the answer is uncomfortable or politically inconvenient.

The deeper mission, however, is for Grok to genuinely understand the nature of the universe. There are many known-unknowns and infinite unknown-unknowns. If Grok can help us understand enough to ask the right questions that lead us to more answers, it will have served its purpose. For the brief period before it helps us with that, every other SpaceXAI product sits on top of Grok. MACROHARD's coding agents run Grok behind the editor, Grokipedia is written by Grok, and Imagine's multimodal image and video generation uses the Grok model family.

Maximally truth-seeking is absolutely essential to ensuring a good AI future for humanity

— Elon Musk (@elonmusk) July 25, 2025

Grokipedia

Grokipedia is the real-time, artificial intelligence-generated alternative to Wikipedia. Articles are continuously written and updated by Grok models, then reviewed and edited by a hybrid network of human contributors and other models. The value is two-fold: First, the human-only Wikipedia is structurally too slow to keep up with how fast the world now changes. Second, the legacy editorial bias can be measured and corrected when the writing layer is a model, where you can audit and tune the prompt rather than argue with thousands of volunteer editors.

Grokipedia fits Elon's broader pattern of pushing accuracy and transparency across his platforms, from the introduction of Community Notes on X to the open-sourcing of the X recommendation algorithm, now applied to the AI-generated, dynamic encyclopedia.

Imagine

Imagine is SpaceXAI's image and video generation product, showcasing the wide array of capabilities from the Grok model family. Users describe what they want in plain language and Imagine returns photo-realistic images or short videos that can seamlessly be stitched together to create long videos. The product competes most directly with Midjourney, OpenAI's former Sora, and Google's Nano Banana Pro 2 + Veo 3.1 video generation efforts. Imagine is also the visual engine behind the next wave of content inside X, including generated thumbnails, profile artwork, and the growing video pipeline.

The future of AI is primarily video understanding and generation, because photons are by far the highest bandwidth form of communication. These are essential tools for AGI.

Worth mentioning that Imagine is positive gross margin for @xAI, not a money loser.

— Elon Musk (@elonmusk) March 30, 2026

Elon made the statement shortly after OpenAI announced on March 24, 2026 that it was shutting down Sora, its own consumer video generation product, citing weak unit economics. The Head of Sora, even admitted: "the economics are currently completely unsustainable" in October 2025.

Where OpenAI saw a money-losing consumer product worth retiring, Imagine remains positive gross margin, meaning on average, every generation produced earns more than it costs to run. That gives SpaceXAI both a technical lead, since it is running the workload competitors are stepping back from, and a financial cushion most consumer artificial intelligence products have not been able to build.

X

X, formerly Twitter, is the necessary data pipeline that feeds the SpaceXAI portfolio, but the seemingly new vision behind it is older than the company itself. In July 2000, Elon and David Sacks wrote an internal product roadmap for x.com and PayPal that mapped out how to turn the original X into what Elon described as "the most valuable national institution in the world," by combining banking, brokerage, insurance, payments, lending, and high-yield accounts into a single seamless online platform.

When eBay acquired PayPal, the plan was never executed, and parts of it were even rolled back. Elon revived the document at an October 2023 internal all-hands meeting after the Twitter rebrand, walking his team through the 23-year-old roadmap and tying it directly to his everything-app vision for X. The stated goal is to turn X into "The Everything App," a single application where every part of a person's digital life happens, modeled on China's 'WeChat.'

By building in-app banking and brokerage accounts, users can hold cash, send money, and trade equities without leaving the platform. They're also building peer-to-peer payments at WeChat scale, replacing Venmo and Cash App for everyday transactions, a dating product designed to displace the Match Group portfolio holding popular dating apps: Tinder and Hinge, a job search and recruiting platform that competes head-on with LinkedIn, creator monetization tools that let writers, video producers, and podcasters earn a living directly from advertising and subscription revenue, replacing and reducing the middle layer that YouTube and Substack currently take.

TERAFAB: The Tesla - SpaceXAI Chip Factory

TERAFAB is the gargantuan (4 mile x 1 mile), end-to-end, $55 billion custom, ultra-high-performance AI chip (logic + memory + advanced packaging) fabrication facility that will be built by Tesla in partnership with SpaceXAI and Intel. In March 2026, Elon stated: "TERAFAB will technically be two fabs, each making only one chip design."

TERAFAB exists because even if all chip fabs in the world accelerate their production, the world would continue to operate at around 20% of the needed capacity for years. In the official TERAFAB announcement, Elon stated: "Goal is a trillion watts of compute/year" and followed it up with: "Most must necessarily go to space, as US electricity is only 0.5TW."

Three of the company's most important roadmaps: Starlink Generation 3, Tesla's Robotaxi and Optimus silicon, and xAI's training compute, all depend on custom chips that today come almost entirely from Taiwan Semiconductor Manufacturing Company (TSMC). That single-supplier dependency is not only risky for achieving the actual mission, but it's the kind of geopolitical exposure a public market does not reward. Building the fab in the United States will remove the dependency and bring strategic silicon onshore, which is the kind of structural advantage (vertical integration) that compounds over time.

That dependency is exactly why, in April 2026, Intel joined TERAFAB as the primary foundry partner. Intel is contributing its leading-edge 14A and 18A process nodes alongside the high-volume manufacturing capacity Elon's companies will need to absorb: surging demand from Tesla robotics, Starlink Generation 3 silicon, and xAI training compute. The deal was sealed publicly at Intel headquarters with photos of Elon and Intel CEO Lip-Bu Tan shaking hands, and it is one of the strongest validations to date of Tan's foundry-first turnaround at Intel.
‍

pic.twitter.com/CrThu6GaBY

— Elon Musk (@elonmusk) May 8, 2026

After visiting Intel's Oregon fab in early May 2026, Elon posted that it was an honor to tour the facility and that he was looking forward to a great partnership between Intel, SpaceX, and Tesla. On Intel's most recent earnings call, Tan returned the compliment, saying he could think of no better partner than Elon to explore unconventional ways to improve chip manufacturing and that Intel expects to learn at least as much from Elon's process-optimization expertise as it shares from its own.

Elon framed the TERAFAB rationale in his own words in a recent X post, calling out the strategic case for vertically integrated silicon and pointing at the workloads the facility is sized to serve. Though there are several clear benefits to controlling their own destiny and building TERAFAB, Elon has explicitly stated the main reason for the decision is simply limited supply. Even if every chip manufacturing facility produced as much as they're willing to, there would still not nearly be enough chips in the world.

How SpaceX Makes Money Today

SpaceX makes money in two main ways. Launch revenue comes from selling rides to orbit on Falcon 9, primarily to NASA, the United States Department of War, and commercial satellite operators. Recurring revenue comes from Starlink subscriptions, where homes, businesses, ships, planes, and an expanding base of mobile users pay monthly fees for high-speed satellite internet. Starlink has now overtaken launch as the larger revenue contributor, which is why the charts that follow focus on the subscriber and revenue compounding curves.

Starlink's subscriber count has compounded annually, and the trajectory from beta launch in early 2020 to 10 million paying customers in six years is what justifies treating SpaceX as a consumer connectivity business rather than just a launch provider.

SpaceX has publicly targeted 25 million subscribers by year-end 2026, and analyst forecasts extending through 2036 range from a conservative case of roughly 100 million to a bull case of 400 million globally, depending on how quickly the Generation 2 constellation expands capacity and how rapidly Direct-to-Cell mobile service contributes. The shaded fan in the chart below visualizes that conservative-to-bull range alongside the established historic trajectory.

The stacked bars below show how Starlink, in mint green, has gone from contributing essentially nothing in 2020 to becoming the majority of total SpaceX revenue by 2025. For context, the recent Anthropic deal announced May 6, 2026 (more discussion to follow) produces more revenue than Starlink generated in 2025. The launch services and other-government lines, in purple, are not shrinking. They are growing on an absolute basis while Starlink scales faster on top of them.

Looking at the same period as a percentage breakdown clarifies what the public-market thesis is actually buying. By 2025, Starlink represented 61 percent of total SpaceX revenue, with launch services declining from 40 percent to 25 percent over two years even as launch volume grew on an absolute basis. The crossover happened in 2024, when Starlink became larger than every other line combined and the company turned, in financial terms, from a launch business with a satellite side project into a satellite business with a launch side capability.

The IPO Deal: Size, Schedule, and Structure

SpaceX is going public for two reasons. The first is structural readiness. After twenty-four years of building, the company has the profitability, the launch cadence, and the recurring satellite revenue to support a public listing without quarterly pressure compromising long-horizon work.

The second reason is more important, and it explains why the listing is happening now rather than five years from now: the scale of what comes next exceeds what private capital can fund alone. Starship V3, the Starlink V3 constellation, point-to-point Earth transport, a manned Mars architecture, the worlds largest end-to-end chip manufacturing facility, mass drivers on the moon, and orbital data centers: none of those line items can be sustained by secondary tender offers and SpaceX bond issuances.

SpaceX's valuation has compounded across many private financing rounds and secondary tender offers between 2015 and 2025 (the main points are plotted below), with each round pricing the company higher than the previous. The log-scale chart below normalizes the visual so each tenfold step covers equal vertical space, which makes the steady multi-decade climb easier to read than a linear plot would suggest. Even so, there's clearly been a recent acceleration over the past few years, as the increasing slope on the log scale plot indicates:

The linear scale chart is below:

Per-employee valuation is one of the cleanest measures of how much economic value a company creates per worker on payroll. At the $1.75 trillion target with roughly 15,000 employees, SpaceX comes out at about $117 million per employee. That translates to roughly four times the per-employee leverage of the most efficient Big Tech firms and forty times what legacy aerospace contractors typically achieve. Put differently, every SpaceX employee carries the economic load of an entire small public company.

The IPO deal was reportedly internally codenamed Project Apex. The underwriting syndicate includes at least 21 banks, with Goldman Sachs, Morgan Stanley, and Bank of America leading the way.

Before the doors open to retail, the company executed a five-for-one stock split. Pre-split, a single SpaceX share carried a $526 price tag. Post-split, that price drops to ~$105/ share at a $1.25 trillion valuation ($147/ share at a $1.75 trillion valuation).

Insiders will hold the vast majority of equity through the full 180-day post-IPO lockup, which on a June 12 listing expires roughly December 15. When a man suggested Elon could roll a staggering post-lockup capital gain into a Qualified Opportunity Fund tied to a future TERAFAB semiconductor site, then lease the asset back to SpaceX itself, Elon publicly closed the door by reassuring he is not interested in some elaborate scheme for deferring capital gains:

At the nearly $2 trillion target valuation, SpaceX trades at high multiples relative to public-market peers. The Price-to-Sales and Enterprise-Value-to-EBITDA ratios shown below assume 2025 financials of roughly $18.7 billion in revenue and $8 billion in Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA). The multiples are defensible if the long-term Starlink, Starship, and other projects grow as projected. If all the projects come to fruition at the desired pace, the valuation could be considered too low.

The company is going public, but Elon is not relinquishing control. SpaceX will use dual-class shares so that Elon's economic stake of around 42 percent post-listing translates into roughly 85.1 percent combined voting power. This is the same structural choice Meta and Google made when they listed, and it ensures the long-term Mars mission cannot be derailed by quarterly activist pressure on the share price. The chart below visualizes the gap between economic ownership and voting power that the dual-class structure creates.

Elon's intent here was expressed in a recent X exchange: a post summarized reported SpaceX board developments: a guarantee that Elon cannot be fired, paired with a trillion-dollar pay package tied to building a Mars colony. Elon's reply in the same thread explained the rationale:

"Yes, I need to make sure SpaceX stays focused on making life multiplanetary and extending consciousness to the stars, not pandering to someone's bullshit quarterly earnings bonus! Obviously, IF SpaceX succeeds in this absurdly difficult goal, it will be worth many orders of magnitude more than the economy of Earth, but don't expect entirely smooth sailing along the way."

The thread is embedded here:

Yes, I need to make sure SpaceX stays focused on making life multiplanetary and extending consciousness to the stars, not pandering to someone's bullshit quarterly earnings bonus!

Obviously, IF SpaceX succeeds in this absurdly difficult goal, it will be worth many orders of magnitude more than the economy of Earth, but don't expect entirely smooth sailing along the way.

— Elon Musk (@elonmusk) May 15, 2026

SpaceX's S-1 reveals a staggered insider lockup rather than one large unlock. Up to 20 percent of eligible shares release after Q2 2026 earnings, plus another 10 percent if the stock trades 30 percent above the IPO price for 5 of any 10 days. Smaller 7 percent tranches free up at 70, 90, 105, 120, and 135 days post-IPO, then another 28 percent unlocks after Q3 2026 earnings, with the remainder releasing at the 180-day mark. Elon is excluded from the early-release provisions, so his shares stay locked the longest.

Starship: Changing Launch Economics

From the Merlin engine that nearly bankrupted the company in 2008 to the Falcon 9 fleet that now flies more than a hundred times per year, SpaceX's path has been a 24-year exercise in stubborn iteration and relentlessness. Each generation of vehicle taught the company something nobody else in launch had bothered to learn at scale: how to land an orbital booster, how to refly it within weeks, how to drop launch cost by an order of magnitude.

The annualized launch cadence today is greater than the combined output of every other launch provider on Earth, and Starship, the fully reusable super-heavy vehicle now flying out of Boca Chica and Cape Canaveral, is on track to make Falcon 9 economics look tiny.

Launch cadence is the operational health metric for the rocket business. SpaceX flew 165 missions in 2025, more than the rest of the world's commercial launchers combined for most of the decade. The 2025 share figure dropped not because SpaceX slowed but because China's launch ramp accelerated dramatically in the same window. That being said, on absolute volume, SpaceX still flew more orbital missions in 2025 than any country has ever flown in a single year.

The chart below stitches together what SpaceX and the rest of the world have actually launched to orbit each year since 2018, then sketches where that trend could go through the end of the decade. The bars for 2018 through 2025 are reported actuals; the bars for 2026 through 2030 are projections. Keep in mind, this data is shown on a log scale:

To make those bars feel like real cargo rather than abstract tonnes, it helps to picture what one Starship actually carries. The upper-stage spacecraft is designed to lift more than 100 metric tons of payload to low Earth orbit in fully reusable mode (roughly 110 U.S. tons, or, if a more vivid unit helps, nearly two million hamsters).

With orbital refueling from dedicated tanker Starships topping the main vehicle off after it reaches orbit, that same upper stage can carry several hundred metric tons of cargo all the way to the Martian surface, because tankers refilling the ship in low Earth orbit are what give it the propellant budget for the full transfer burn.

For comparison, the Saturn V that carried Apollo to the Moon delivered about 45 tons to lunar trajectory, and Perseverance, the rover that has been driving around Mars since 2021, weighed about one ton at landing. A single fully loaded Starship arriving at the Martian surface is therefore the rough equivalent of several hundred Perseverances, or many complete habitat modules, or the heavy industrial equipment needed to start extracting water from ice, breathing oxygen out of the atmosphere, and printing structures out of Martian regolith. Five to ten Starships in a single transfer window deliver roughly the entire mass of the International Space Station, which took more than two decades and dozens of launches to assemble in Earth orbit.

That is where the upper bars on the chart actually matter. A handful of Starships in 2030 is enough to plant a research outpost. Hundreds of Starships across consecutive transfer windows is enough to build a working base. Reaching the thousand-Starships-per-year cadence Elon describes is the scale at which Mars stops being a science station and starts becoming a habitable city, with the cumulative mass to deliver power grids, food production domes, fuel plants, on-site factories, and supplies for tens of thousands of residents.

Before SpaceX figured out how to reuse rocket boosters, sending one kilogram of cargo to low Earth orbit cost around ten thousand dollars. That meant spaceflight was mostly only accessible to national budgets. By 2025, with the booster routinely flying itself back, landing on a drone ship or land pad, and getting refurbished for another mission, the cost per kilogram had fallen below one thousand dollars.

The same launch business that used to manage a handful of flights per year now flies one hundred and sixty-five missions annually. Reusability is the engineering breakthrough behind every other number on this page, because every Starlink satellite, every Department of War payload, and every commercial mission depends on getting the booster back and flying it again.

Two related numbers tell the constellation story. Total satellites launched, the dashed line in the chart below, is the cumulative count SpaceX has put on rockets since 2019. Satellites still in orbit, the solid line, is smaller because older Generation 1 satellites have been deorbited as they reach their roughly five-year design end of life. The gap between the two lines widens after 2024 as the first large launch cohorts begin retiring; this is by design, not a sign of failure.

Raptor is the methane-and-oxygen engine that powers both the Starship upper stage and the Super Heavy booster. Each Starship stack uses 39 Raptor engines, so reaching a Starship-per-day launch cadence eventually requires Raptor production at roughly 14,000 engines per year. The production rate has climbed from a few dozen per month in 2024 to roughly 200 per month by 2026, and the McGregor and Hawthorne facilities are scaling further to meet long-term demand from both Starlink V3 and the Mars fleet.

Production rate of complete Starship upper stages and Super Heavy boosters is the other half of the Starship economics equation. Past years averaged 5 to 15 vehicles total across both classes, heavily weighted toward iterative prototypes that were often scrapped after structural and cryogenic tests. The 2026 inflection comes from the Starfactory facility at Starbase, which Elon has stated is designed for 1,000 ships per year at full rate. The 10,000 per year ceiling mentioned in his January 2026 comments is aspirational, but it is the production scale a true Mars-fleet timeline would require.

Starlink: Why Connectivity Matters

SpaceX runs the largest commercial satellite constellation in history, the most active orbital launch operation on the planet, and a growing in-space services arm. Starlink alone has crossed 10 million subscribers across 105 countries. Launch revenue, anchored by NASA, the US Department of War, and a deepening commercial customer base, complements that subscription engine and contributed roughly $4.2 billion in 2025.

Average Revenue Per User (ARPU), is the consumer-business health metric for Starlink. ARPU has compressed from approximately $95 per month in 2023 to roughly $81 per month by 2025, which is the expected pattern for any consumer service expanding into emerging markets where willingness to pay is expectedly lower than in the United States or Western Europe. Total revenue grew despite the ARPU decline, which is the unit-economics signal investors look for: per-user revenue dropping while aggregate revenue accelerates means the subscriber base is compounding faster than per-user pricing is eroding.

Starlink's Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) nearly doubled in just one year from $3.8B in 2024 to $7.2B in 2025. Margins of this level (shown below), are typical of mature software businesses, not satellite network operators, and they are what made the Starlink line the most valuable single asset in the company as of 2025. Once a Starlink satellite is in orbit and a user has a dish installed, the incremental cost of serving that user is close to zero, which is why the operating leverage looks so attractive at scale.

Buried inside the Starlink line item is Starshield, the classified military and government version of the constellation built specifically for the Department of War (DoW), the National Reconnaissance Office (NRO), and allied intelligence services.

Starshield-equipped spacecraft carry classified payloads, run encrypted communications for the warfighter, and host imaging and signals intelligence sensors that Starlink commercial customers never see. Most contract values are not public, but the segment is already a billion-dollar business on disclosed work alone, and analyst expectations have it scaling to multi-billions by 2028 as crewed and unmanned national security workloads migrate to space.

The Numbers: Valuation, Multiples, and Scenarios

With the S-1 coming public on May 20th, a global roadshow led by Goldman Sachs, Morgan Stanley, Bank of America, JPMorgan, and Citigroup will begin.

On the institutional side, BlackRock has reportedly discussed purchasing between $5 - $10 billion in the offering. A commitment at the top of that range would amount to roughly thirteen percent of the entire $75 billion raise from a single asset manager, which is the kind of vote of confidence that could cause other institutions to tail along. It would be one of the largest single institutional bets on a tech IPO ever placed.

On the retail side, the deal is engineered to feel accessible. The five-for-one split is the obvious lever. The directed share program is the less obvious one. Retail allocations are expected to flow through the major US brokerages, with Robinhood, Fidelity, and Schwab participating in a program that reserves a portion of shares for Starlink customers and the immediate family members of SpaceX employees.

Beyond that, Elon has been clear, going back years, about who gets first call when one of his companies goes public. In June 2024 he posted: "I've mentioned something like this before, but, if any of my companies goes public, we will prioritize other longtime shareholders of my other companies, including Tesla. Loyalty deserves loyalty." SpaceX has not formally confirmed the mechanism for routing Tesla-shareholder priority into the $SPCX book.

With the headline numbers public, the question every analyst will press the underwriters on is how the proceeds are actually deployed. Based on prospectus reporting and the company's stated priorities, the rough split looks like this: roughly 40 to 45 percent toward the Starship program and Mars architecture, around 28 to 30 percent toward Starlink V3 expansion, 15 to 20 percent toward xAI synergy and orbital data center infrastructure, and the balance toward general corporate purposes.

Here is some context for what has happened the day after a record IPO. Saudi Aramco's listing in late 2019 saw a modest first-day pop and then years of range-bound trading. Alibaba's 2014 listing surged on day one and continued higher for years before regulatory pressure cooled it. SpaceX has the unusual combination of an unprecedented raise and an exceptional growth thesis, which makes the trajectory hard to map onto either precedent. Plus, any guess here would be using reasoning by analogy, which is not something Elon would adovocate for.

Combining historic actuals with three projection scenarios shows where the public-market growth story could go. The historic line through 2025 is well-documented from prospectus excerpts and Reuters reporting; the projection fan from 2026 to 2030 reflects analyst consensus estimates. The bull case reaches $180 billion by 2030; the base case $120 billion; and even the conservative scenario would make SpaceX one of the fastest-growing large-cap companies in modern history.

Revenue is one read of the future, but the question public-market investors care most about is market cap. Scenario-mapping based on analyst guidance and milestone delivery puts the conservative case at roughly $2T by 2027 and $3T by 2030, a base case at $2.5T and $5T, and a bull case at $3T and $8T. The fan below visualizes the spread.

The Total Addressable Market (TAM) for just Starlink stretches across three distinct customer segments. The residential segment is the largest by far: roughly three billion people globally live without broadband connectivity today, and Starlink is the only constellation that can reach them economically without the trenching and tower infrastructure terrestrial providers require. Maritime and aviation are smaller in raw user count but higher-margin, and enterprise and defense are sticky, contract-anchored revenue with multi-year commitments.

Starship is designed to make Falcon 9 economics look expensive. The target cost per kilogram to low Earth orbit (LEO) drops from approximately $2,000 with the current Falcon 9 fleet to as low as $10 with a fully reusable Starship operating at the cadence Elon has described. The payload capacity also jumps from roughly 22 metric tons per launch to between 100 and 150 metric tons. These numbers, if achieved at production scale, do not merely change SpaceX's economics; they restructure the entire launch industry and unlock workloads that are currently not economic.

The roadmap from this IPO to a self-sustaining Mars colony is the primary reason for the offering. Public capital funds Starship development through 2028, the first uncrewed Mars landings in the 2028 to 2030 window, the crewed missions through the 2030s, and the long-term city/ colony buildout across the 2040s and 2050s. Each milestone is a budget line in the prospectus rather than a slide-deck aspiration, and Elon's control structure exists specifically to keep the company aimed at that decade-long horizon when quarterly earnings pressure could otherwise pull it elsewhere.

What are Orbital Data Centers?

Orbital data centers are server farms placed in space rather than on the ground. The architecture solves four constraints terrestrial data centers cannot:
1. Solar power in low Earth orbit is continuous and free of grid limits.
2. Cooling is radiation directly into the 2.7-Kelvin background of space at zero marginal cost.
3. Latency to any user on Earth is a handful of milliseconds through Starlink's optical inter-satellite links.
4. Capital cost, the constraint that historically made the idea absurd, collapses under Starship's economics.

The S-1 also disclosed a major cloud-compute partnership with Anthropic, the maker of Claude. Signed in May 2026, the deal gives Anthropic exclusive access to the full capacity of SpaceX's Colossus I and Colossus II data centers in Tennessee and Mississippi, two facilities originally built by xAI for Grok training that together carry over one gigawatt of power and hundreds of thousands of GPUs. Anthropic pays $1.25 billion per month from July 2026 through May 2029, roughly $15 billion a year and up to about $45 billion over the three-year term, with either side allowed to terminate on 90 days' notice. The arrangement turns SpaceX into a "neocloud" operator that monetizes capacity beyond its own internal needs, adds a revenue stream that could rival the company's entire 2025 top line of $18.7 billion, and signals a clear strategic pivot toward AI infrastructure heading into the Nasdaq debut.

In a recent Forbes interview on May 19, 2026, Elon said: "Data centers in space is much easier than people may think." SpaceX's S-1 explicitly breaks out three main operating segments with separate revenue lines:
1. Space segment (launch services, spacecraft, etc.)
2. Connectivity segment (Starlink, etc.)
3. AI segment (this is where orbital compute shows up)

The next-decade thesis goes further than Mars. Asked recently where AI compute would settle once the energy and cooling math is run end to end, Elon answered in January 2026: "The lowest cost place to put AI will be space, and that'll be true within 2 years, maybe 3 ... 3 at the latest."

If the trajectory holds, the most expensive constraint on AI compute today: gigawatt-scale energy on the ground, stops being the constraint of the buildout. The whole thing moves above the atmosphere, and the pipeline that moves it there is SpaceX's alone.

Why a Mass Driver on the Moon?

A lunar mass driver is an electromagnetic catapult on the surface of the Moon that flings raw material to orbit. It's literally a fixed track that flings payloads at orbital speed without the need to burn any fuel at all. The Moon's gravity is roughly one-sixth Earth's and there is no atmosphere to fight, so launching mass off the lunar surface takes about thirty times less energy than launching from Earth. The strategic logic flows directly from that math. The reason that matters for SpaceX is that going to Mars and beyond is bottlenecked by how much propellant and how much mass you can lift off Earth.

If Starship is going to build Mars-bound vehicles in space at scale, the raw material has to come from somewhere cheaper than the Earth's surface. The Moon, with its anorthositic crust, polar ice, and oxygen-rich regolith, is that cheaper source.

Some Risk Factors

SpaceX's Form S-1 details a deep list of risk factors, most centered on execution at unprecedented scale. Any delay or failure in developing Starship at the required cadence, reusability, and capabilities would directly slow the next-generation satellites, satellite-to-mobile connectivity, and orbital AI compute that the entire growth story depends on.

Launch failures, test-flight anomalies, satellite malfunctions, space debris and collision risks, ground station and manufacturing outages, and supply chain disruptions are all called out as inherent and material. The AI compute backbone also depends on uninterrupted access to power and specialized chips the company does not yet make in-house at the volume needed. Some additional regulatory and geopolitical risks are outlined here:

The regulatory and competitive picture is similarly heavy. Launch and reentry licenses with the Federal Aviation Administration (FAA), spectrum and satellite authorizations with the Federal Communications Commission (FCC) and foreign regulators, and evolving AI laws across the US and abroad can each delay or constrain operations, with active investigations already noted on Grok's content moderation. Cybersecurity threats, talent retention in a fierce engineering market, competition from new Chinese and European launchers and from emerging AI players, and macroeconomic or geopolitical instability are all flagged. The AI segment specifically is described as recently formed and rapidly evolving, with integration and execution risks that have no clean historical precedent.

The financial and governance picture carries its own risks. SpaceX is targeting massive capital spending across Starship, TERAFAB, satellites, and orbital data centers, while recent losses run into the billions even as Starlink scales. Total addressable market estimates may turn out wrong, and longer-dated initiatives like orbital AI compute, in-house chip manufacturing, the lunar economy, human augmentation, and Moon/Mars transport involve unproven or nonexistent technology with no guarantee of commercial viability. The S-1 also flags key-person risk: Elon will remain CEO, CTO, and Chairman with majority voting control through Class B shares, the company is electing controlled-company status, and active xAI-related litigation estimated at around $530 million is ongoing.

What Owning a Piece of SpaceX Means

For the first time in history, anyone with a brokerage account can own a piece of the company most likely to put humanity on the Moon, on Mars, and on the planets beyond. That is what this IPO really is. Sure it is great to participate in a historic financial event, but supporting and participating in the most ambitious civilizational project of our lifetimes is something generations to come will appreciate.

My Thoughts

This is an opportunity millions have dreamed of for years. Oh-so-many want to invest in SpaceX and just get more exposure to any Elon-company. As AI advancements continue to accelerate, there is no one on planet Earth I'd want in charge of the company I own a piece of than Elon Musk. He and the teams are working 24/7 to make progress as fast as possible on the projects and initiatives that no other team dare have the courage to tackle. The core advantage of having that great team is phenomenal decision-making combined with rapid iteration. Elon often says that pace of innovation/ speed is the best offense and defense. By investing in SpaceX, that is what you are getting- the most rapidly improving and most efficiently designed AI team on the planet.
‍

In my view, this is the crux of what it means to be a good investor. A good investor is simply assessing talent- is the current team excellent, and do they have the abilities to build an even better team in the future? If yes, incredible progress will follow, as demonstrated with the evolution of the Raptor engines from version 1 to version 2 to version 3:
‍

If Elon is right, as he usually is- SpaceX will be the only company that not only has the courage to build data centers in space, but also the ability to build them. They have all the pieces necessary for long-term success.

We have a short window of life that happens to overlap with the moment our species takes its first real steps off Earth. The title of this article is not an accident. Owning a piece of SpaceX, at this moment in time, is owning a piece of history. The choice is whether to be a spectator or a participant. Let's make the future great!

The author may own stock in SpaceX and/ or Tesla. This article is not investment advice.

Sources

Sources