In the span of five days, SpaceX filed for what could be the largest IPO in history, launched and landed the most powerful rocket ever built in its new V3 form, and revealed for the first time just how much money — and how many losses — sit beneath the surface of Elon Musk's empire. This is not a normal week for a space company.

01 — Starship V3: Flight 12 Debrief

On May 22, 2026, SpaceX launched Starship Flight 12 from Starbase, Texas — the first flight of the entirely redesigned Starship Version 3, and the first launch from the new Pad 2. After a scrubbed attempt on May 21 due to pressure issues in the Ship quick disconnect, the rocket lifted off at 10:30 p.m. local time and did not disappoint.

The V3 architecture brings sweeping changes: Raptor 3 engines with higher thrust and simplified design, a novel fuel system in the booster, only three larger grid fins replacing the previous six, an integrated hot-staging system removing the expendable interstage, and significantly increased propellant capacity. SpaceX also tested intentional heat shield tile removal to study cascading damage — the kind of data that makes future crewed missions safer.

02 — The S-1 Drops: SpaceX Goes Public

On May 20 — one day before Flight 12's first attempt — SpaceX publicly filed its S-1 registration statement with the SEC. For a company that had kept its finances entirely private since its founding in 2002, this was a seismic disclosure. The numbers inside are both impressive and complicated.

The Starlink connectivity segment alone generated $3.26 billion in Q1 2026 — profitable at an operating level of $1.19 billion. The picture for the AI segment, however, is starkly different: SpaceX's newly acquired AI division (incorporating xAI and rebranded "SpaceXAI") posted a $2.47 billion operating loss in Q1 2026 alone, and a $6.36 billion loss for all of 2025.

Elon Musk controls 85.1% of combined voting power through Class B shares, meaning minority shareholders will have virtually no say in corporate direction. The IPO, targeted for listing as early as June 12, 2026 on a yet-to-be-confirmed exchange, would raise an estimated $30 billion.

03 — The Moon Mandate:
Artemis, Starship, and the Race to the South Pole

To understand why Flight 12 carried weight far beyond a test flight, you need to understand NASA's Artemis program — and precisely how much of it depends on Starship working.

In 2021, NASA awarded SpaceX a $4.05 billion contract to develop two human-rated Starship vehicles as its Human Landing System (HLS) — the craft that will carry astronauts from lunar orbit down to the surface of the Moon. The prize: delivering the first woman and the first person of color to walk on the Moon, at the lunar South Pole, a region that harbors permanently shadowed craters believed to contain billions of tons of water ice. That ice is not just scientifically significant. It is rocket propellant. It is oxygen. It is the foundation of any future permanent lunar base.

The critical dependency — the one that links Flight 12 directly to astronauts walking on the Moon — is orbital refueling. Starship HLS cannot carry enough propellant from Earth to reach the Moon's surface and return. The solution: build a propellant depot in low Earth orbit, fill it with up to 10 or more tanker launches of liquid methane and liquid oxygen, then transfer that fuel to the HLS Starship before it departs for the Moon.

This "gas station in space" architecture is unlike anything ever attempted at scale. SpaceX is targeting June 2026 for the first integrated orbital refueling demonstration — two Starships docking belly-to-belly in low Earth orbit, one transferring cryogenic propellant to the other. A "target" Starship launches first. Three to four weeks later, a "chaser" Starship rendezvouses, docks, and transfers fuel. Then both deorbit.

If that demonstration succeeds, an uncrewed Starship lunar landing — a full dress rehearsal at the Moon — is targeted for June 2027. Only after that can NASA authorize astronauts to descend to the surface. The earliest a crewed Moon landing could occur under SpaceX's own internal timeline: September 2028.

Flight 12 checked none of those boxes directly. What it did prove is that Starship V3's propulsion, thermal protection, and avionics architecture is fundamentally sound enough to fly — and survive — without all engines firing perfectly. That is the foundation. Everything else has to be built on top of it, and the clock is running.

04 — The Mars Equation:
Musk's 50/50 Bet on 2026

The Moon, for Elon Musk, is not the destination. It is a waypoint. The destination — the one he has described as the core reason SpaceX exists — is Mars. And the 2026 Earth-Mars transfer window, which opens in November and closes in December, represents the next opportunity to try.

In a video posted to X in mid-2025, Musk gave his company a 50-50 chance of launching an uncrewed Starship to Mars by late 2026. He was explicit: the entire timeline hinges on whether Starship can master orbital refueling. Without propellant transfer in Earth orbit, a Mars trajectory is impossible — the fuel required simply cannot be launched in a single vehicle.

There is a tension embedded in this timeline that the S-1 filing does not resolve. SpaceX's internal documents — reported by the Wall Street Journal — suggested the company had privately told investors it would "prioritize going to the Moon first" and attempt Mars at a later time, effectively stepping back from the aggressive 2026 Mars target. Publicly, Musk has not formally abandoned the 2026 window. Privately, the engineering dependency is clear: orbital refueling must be proven before either the Moon landing or a Mars departure is attempted.

If the orbital refueling demo planned for June 2026 succeeds, the 2026 Mars window becomes theoretically plausible. If it slips — as many milestones in Starship's development have before — the next Earth-Mars alignment is late 2028. Musk himself acknowledged: "If Starship were not ready by that time, SpaceX would wait another two years before trying again."

What makes 2026 uniquely high-stakes is the convergence: the IPO, the orbital refueling demo, the potential Mars window, and the Artemis schedule pressure all fall within the same twelve-month band. SpaceX has never had this much riding on a single year's performance. Flight 12 was the opening shot.

05 — Starlink: The Machine
That Funds Everything

To understand how SpaceX can simultaneously fund Starship development, an AI division burning billions, and a Mars program, you have to understand Starlink — and how radically its economics have changed in three years.

In 2022, Starlink was a promising but money-losing satellite internet service with a few hundred thousand subscribers and significant hardware subsidies per terminal. By the end of Q1 2026, it had 10.3 million subscribers across more than 100 countries, generated $3.26 billion in a single quarter, and operated at a 63% adjusted EBITDA margin — comparable to the best software businesses in the world, running on a constellation of nearly 10,000 satellites in orbit.

Starlink works because SpaceX is its own launch provider. Every Falcon 9 that lifts a batch of Starlink satellites costs a fraction of what a competitor would pay for equivalent capacity — because the rocket, the launch pad, and the integration are all internal. Reusability compounds this advantage: a Falcon 9 first stage has now flown more than 25 times in some cases, amortizing the manufacturing cost across dozens of missions. The more Starlink grows, the more launch capacity SpaceX needs. The more launch capacity SpaceX runs, the cheaper each unit becomes.

This flywheel is what makes SpaceX's IPO valuation simultaneously defensible and extraordinary. A $1.5 trillion target implies investors are pricing not just Starlink's current cash flows, but its trajectory: projected revenue of $22–24 billion in 2026, with Starlink expected to account for roughly 79% of total company revenue as Starship development costs drag the Space segment into operating losses.

06 — The AI Wildcard:
SpaceXAI and the $15B Anthropic Bet

The most surprising disclosure in SpaceX's S-1 was not the Starlink numbers — analysts had modeled those for years. It was the AI segment. SpaceX's acquisition of xAI (Elon Musk's artificial intelligence company, maker of the Grok chatbot) and its rebranding as "SpaceXAI" revealed a business that is simultaneously massive in ambition and deeply in the red.

The AI segment posted a $6.36 billion operating loss in 2025 and a further $2.47 billion loss in Q1 2026 alone. Yet the S-1 frames SpaceX primarily as an AI company — the filing uses AI-related terms in 47% of segment-specific language and allocates 93% of the company's stated $28.5 trillion TAM to AI-related markets.

The strategic logic is coherent even if the losses are eye-watering: Starlink provides global low-latency connectivity. Orbital compute satellites (deployment targeted 2028) provide the processing layer. xAI / SpaceXAI provides the model layer. In Musk's vision, SpaceX becomes the infrastructure backbone of a planetary AI network — one that operates above national borders and is accessible anywhere a Starlink terminal can see the sky.

Whether that vision justifies the losses depends on execution timelines that no S-1 can guarantee. What the filing makes clear is that Starlink profits are currently subsidizing everything else — Starship, SpaceXAI, Mars ambitions, and the $20 billion bridge loan that comes due 15 months after the IPO.

07 — A Week in History: Timeline

08 — What Could Go Wrong:
The Honest Risk Register

SpaceX has defied skeptics consistently enough that dismissing its timelines has become an embarrassing habit for analysts. And yet, intellectual honesty requires acknowledging where the cracks could form.

None of these risks are fatal. All of them are real. The SpaceX that filed its S-1 in May 2026 is structurally stronger than any version of the company that came before — Starlink's cash generation provides genuine insulation against near-term volatility. But the margin for timeline error narrows significantly once public shareholders are watching quarterly earnings calls.