|
|
SPCX was priced at $135 on June 11 and ran roughly 50% higher within days to about $2.5 trillion, just below Amazon, on the largest IPO on record ($85.7 billion raised). Moody's, S&P Global and Fitch have privately placed SpaceX at investment grade. SpaceX also signed its first post-IPO deal, a $60 billion all-stock buy of Cursor, and Musk floated roughly $1 trillion in revenue by 2030.
This week's premium analysis rebuilds AI1, SpaceX's first orbital compute satellite design by subsytem. The full analysis is free to read for the next 33 hours. After that it stays open only to Mach33 Premium Research members, who also receive the model and assumptions. [Read it now]
If you haven't already, sign up for our weekly podcast [Sign up] . The next is tomorrow at 11:00 AM, where we'll walk through the SpaceX AI1 satellite architecture with special guest McKenzie Read Sandberg, ex-head thermal engineer for SpaceX during the early generations of Starlink.
|
| Latest Analysis |
|
|
| Jun 17, 2026 |
|
SpaceX AI1: A Subsystem-Level Reconstruction
|
|
SpaceX
|
|
SpaceX disclosed the first specification for AI1, its first orbital compute satellite, in June 2026. We reconstruct the satellite subsystem by subsystem from the published figures and the Mach33 Orbital Data Center model, which had already placed both the first unit and the optimized unit inside the range SpaceX has now confirmed. The reconstruction shows a satellite optimized for a single 150 kW figure. Solar, compute, and radiator are all designed to handle a 150 kW load.
|
|
|
Model
Premium
|
|
AI Subsystem Model 6/19/26
|
|
This workbook puts SpaceX's published AI1 orbital data center to the test, rebuilding the satellite from first principles to see whether its stated...
|
|
|
|
|
|
Read More →
|
|
| Jun 11, 2026 |
|
Five Thousand Futures on IPO Eve: Inside the SpaceX Valuation Model | Episode #7
|
|
SpaceX
|
|
On the eve of the SpaceX IPO, Mach 33 Research walked through the valuation model behind its recently published numbers: a 5,000-trial Monte Carlo spanning roughly 350 assumption ranges. Under the team's preferred EV/EBITDA lens, compute overtakes Starlink as the biggest value driver, with a best-case tail reaching $37.9 trillion by 2040, hinging on ODC adoption, orbital compute demand, multiples, and chip costs. The episode covered how capital allocation flows from Starlink's mid-2030s maturation into AI satellites, why the $75 billion all-primary raise reshaped the model's cash assumptions, and where debt enters through TerraFab and ODC satellite facilities. Audience Q&A touched on demand from humanoid robotics, dawn-dusk orbital mechanics, and SpaceX's gravitational pull on the entire space industry. The team closed by framing the model as a living draft that will keep iterating.
|
|
Read More →
|
|
|
| Industry News |
|
|
|
|
SpaceX agreed on June 16, 2026 to acquire Anysphere, the maker of the Cursor coding agent, for $60 billion in an all-stock deal, its first major acquisition since going public last week, per CNBC. The transaction is expected to close in the third quarter, pending regulatory approval. Earlier reporting had described two possible structures: a roughly $10 billion compute partnership, under which SpaceX would simply supply Cursor the data-center capacity to run its AI, or a full takeover. SpaceX chose the takeover, owning Cursor outright rather than acting as its supplier.
The strategic logic runs through the structure of the AI business, which can be read as a stack of five layers: the electricity that powers the system, the chips that compute, the data centers that house them, the foundation model that does the reasoning, and, at the top, the orchestration layer, which wraps a general AI model into a finished, purpose-built tool for a specific job. Cursor is an orchestration product: it takes a model and turns it into a focused assistant that writes and edits software. A compute partnership would have left SpaceX owning the bottom four layers, energy, chips, infrastructure, and, via xAI's Grok, the model itself, while Cursor kept the orchestration economics at the top. Buying Cursor pulls that final layer in-house, giving SpaceX's AI arm a proven, market-leading product surface rather than a model in search of a distribution channel, and lifting it from a wholesale compute-and-model supplier to an owner of the entire chain from the electricity to the end-user application.
The move mirrors SpaceX's launch playbook, where capacity rented to outside customers during the buildout, as with its recent wholesale data-center lease to Anthropic, has progressively been internalized, the same way launch slots once sold to third parties increasingly went to its own Starlink network as the business scaled. The full payoff is contingent: it is realized only if Grok and Cursor can compete on model and product quality with OpenAI, Anthropic, and Google, a different contest than the one SpaceX has already won in launch.
Read More →
|
|
|
|
SPCX, priced at $135 on June 11, closed its debut at $160.95. It continued to climb about 20% on Monday, June 15, and reached $206.19 by June 16, a market value near $2.73 trillion, per Investing.com and CNBC. Trading stayed deep, roughly 244 million shares on Monday after more than 500 million at the debut. The offering also rewrote the record books: about $75 billion raised, near $85.7 billion after the overallotment, per Phemex, the largest IPO on record.
The rally has pushed the implied valuation from $1.77 trillion at pricing to above $2.5 trillion in days, and the distance between price and fundamentals is now the whole debate.
Read More →
|
|
|
|
Elon Musk posted that SpaceX "might be able to reach approximately $1T revenue in 2030," replying to a post that cited Morgan Stanley's estimates. The figure sits well above the sell-side, where Morgan Stanley models roughly $330 billion and Goldman roughly $474 billion for 2030, and far above SpaceX's roughly $18.7 billion in 2025 revenue (per its IPO filing).
Against our own work, the comment underscores how conservative our base case is by design. The latest iteration of our model reaches about $1.8 trillion in revenue by 2040 and assumes roughly one-fifth of Musk's stated launch-tonnage ambition, on the view that space timelines slip and that the projections should hold without heroic assumptions.
Read More →
|
|
|
|
Bloomberg, citing sources, reports that Moody's, S&P Global, and Fitch have each placed SpaceX at investment grade, communicated privately to investors around the IPO. None of the three has issued a public rating, so this sits as a sourced report rather than a confirmed action. The report arrives with SPCX trading around $200 a share, a market cap near $2.5 trillion as of midday June 17, up roughly 50% from its $135 listing and now among the most valuable public companies in the world, just below Amazon by market value.
An investment-grade rating measures ability to service debt. SpaceX raised roughly $85.7 billion in the IPO, above the $75 billion we had assumed, after underwriters exercised the greenshoe over-allotment (per CNBC). That cash, set alongside a separate base of about $75 billion in contracted revenue from investment-grade counterparties, gives unusual forward visibility through a heavy-spend buildout. CreditSights' Zachary Griffiths framed it plainly: "Negative earnings are not typically associated with an investment-grade company, but nothing about this is typical." CreditSights expects SpaceX to issue investment-grade debt shortly after listing.
Still, the balance sheet is not yet set. Our models suggest near-term funding to build out Terafab, with another wave of debt or equity financing in the early 2030s to accelerate the orbital data center (ODC) AI-satellite rollout. That is the practical reason the ratings matter: a lower borrowing cost compounds across the capital SpaceX still has to raise, and the larger-than-expected IPO haul covers only part of it.
Read More →
|
|
|
|
NASA Administrator Jared Isaacman stated publicly on June 5, 2026 that the agency is evaluating alternative launch vehicles for Blue Origin's Blue Moon Mark 1 cargo lander after the May 28 explosion of a New Glenn rocket during a static fire test at Launch Complex 36 (LC-36) at Cape Canaveral Space Force Station destroyed the transporter-erector and damaged the gantry. Speaking at the CNBC CEO Council Summit, Isaacman said that in terms of heavy-lift options, the agency is likely looking at "Falcon Heavy land," per Gizmodo reporting. The Blue Moon Mark 1 uncrewed cargo lander had been contracted to fly on New Glenn as early as fall 2026 to deliver NASA payloads to the lunar surface. Blue Origin has pledged to resume New Glenn flights before year-end, but one source familiar with pad rebuilds estimated 15 months as a best-case scenario for full pad restoration, per Ars Technica reporting cited by Universe Today.
The LC-36 explosion removes Blue Origin's only operational orbital launch pad, and the Blue Moon lander's 23-foot fairing requirement constrains the substitute vehicle pool to Falcon Heavy or a future New Glenn. NASA had awarded Blue Origin contracts just two days before the explosion, including a fall 2026 Blue Moon Mark 1 demonstration mission and two New Glenn flights to carry lunar terrain vehicles for Artemis III through V. With Starship also still in test-flight status and the Artemis III Earth-orbit rendezvous targeted for 2027, the agency now faces a single-point-of-failure risk on its heavy-lift lander delivery chain. The FAA confirmed the static fire was outside the scope of its licensed activities, meaning no new FAA investigation is required, but Blue Origin must still satisfy its own mishap review before any return-to-flight authorization.
Read More →
|
|
|
| Mach33 |
| The Space Finance Group |
|
|
|
