The amended S-1 is live. The terms are fixed. And the gap between what SpaceX is asking and what at least one major analyst firm reportedly thinks it’s worth is $970 billion wide.
That gap is the story, not the IPO mechanics.
What the Amended S-1 Actually Discloses
According to the amended S-1 filed with the SEC, SpaceX has fixed Class A shares at $135, targeting a $75 billion raise through 555.6 million shares, with Nasdaq trading scheduled for June 12 under ticker SPCX. The implied valuation is $1.75 trillion.
The filing discloses xAI operating losses projected at approximately $10 billion for 2026. That isn’t a surprise. The S-1’s quarterly data showed xAI posted a $2.47 billion operating loss in Q1 2026 on $818 million in revenue. The quarterly burn acceleration has been documented. An annualized run rate from Q1 lands near $10 billion in annual losses, the amended S-1 confirms it.
The filing also discloses $269 million in Tesla Megapack purchases in April 2026, bringing xAI’s total energy infrastructure spending with Tesla to approximately $1 billion since 2024. Related-party transactions at this scale require careful investor scrutiny, xAI’s compute infrastructure is being built partly through a sister company’s product line, creating interconnected financial exposure that’s unusual even by frontier lab standards.
SpaceX has reportedly filed with the FCC for approval to deploy a constellation targeting space-based AI compute. The company projects 100 gigawatts of annual capacity by 2028. That figure is SpaceX’s stated target. It hasn’t been independently assessed, and investors should treat it as a forward-looking claim, not an engineering milestone. The 100 GW projection is the orbital compute thesis, and it’s the single biggest variable in the valuation gap.
The Valuation Tension: $780 Billion Against a $1.75 Trillion Ask
Morningstar has reportedly issued a $780 billion fair value estimate for SpaceX. The reasoning tracks: a company projecting $10 billion in annual losses on an infrastructure thesis that hasn’t yet demonstrated the compute revenue to justify it represents meaningful execution risk. $780 billion is still an enormous company. It implies the Starlink business, the launch business, and a credible path to monetized AI infrastructure, without requiring the orbital compute thesis to hit 100 GW by 2028.
The $1.75 trillion ask requires something more. It requires investors to price in the orbital compute thesis materializing, xAI reaching revenue levels that compress the burn rate on a timeline that doesn’t exhaust cash, and no competitor replicating the Starlink bandwidth advantage before the satellite AI layer is operational.
Neither position is obviously wrong. But the spread between them, nearly $1 trillion, is not a rounding error. It reflects a genuine disagreement about whether the infrastructure build-out thesis is priced into the orbital compute story or has yet to be earned.
The American Federation of Teachers has reportedly petitioned the SEC to apply heightened scrutiny to the offering, citing pension fund exposure concerns. If institutional pension funds are being marketed this offering, the fiduciary calculus of buying at a 2.25x premium to a major analyst’s fair value estimate becomes a governance question, not just an investment one.
xAI Revenue vs. Burn (annualized from Q1 2026)
What to Watch
The Race: Where SpaceX, Anthropic, and OpenAI Stand
SpaceX isn’t the only frontier lab heading to public markets, and June 12 isn’t just a SpaceX event.
All three major frontier labs are on the IPO path. Anthropic filed a confidential S-1 on June 1, following its $65 billion Series H close. Goldman Sachs, JPMorgan, and Morgan Stanley are reportedly advising, with an October 2026 listing window. OpenAI’s public market timeline continues to develop.
The sequencing matters. SpaceX goes first, on June 12. That’s the market’s first opportunity to price a frontier AI lab publicly. The gap between the $135 IPO price and first-day trading establishes a reference multiple that every subsequent frontier lab deal will be benchmarked against.
Consider the comparison: Anthropic has not disclosed its S-1 terms, but prior reporting placed a $65 billion Series H at a valuation near that figure. If SpaceX trades at a 20% discount to its IPO price on June 12, the public market signals a materially different risk appetite for frontier AI equity than if it trades at a 10% premium. Anthropic’s October window, and the banks advising it, will read that signal before fixing its own terms.
OpenAI’s path is further out. But the same dynamic applies. The 2026 AI IPO cycle is sequential, and the sequence’s first data point is June 12.
The xAI Economics: Burn, Megapacks, and the Infrastructure Thesis
The $10 billion projected burn is the number that defines everything else in the xAI investment case.
At $818 million in Q1 revenue, xAI is generating roughly $3.3 billion in annualized revenue against $10 billion in projected losses. That’s a loss ratio of approximately three dollars spent for every dollar earned. For context, that’s aggressive even by frontier lab standards, and the frontier lab standard is already aggressive.
The Tesla Megapack spending is the physical expression of the orbital compute thesis. xAI needs energy infrastructure to run data centers now and, eventually, to power satellite compute systems. The $1 billion in Tesla Megapack spending since 2024 represents genuine capital commitment, not vaporware planning. It also represents genuine capital consumption.
The orbital compute thesis, 1 million satellites, 100 GW by 2028, is either the most consequential AI infrastructure bet in history or an extraordinarily ambitious projection that compresses into a more modest outcome. The $970 billion valuation gap between Morningstar’s estimate and the IPO target is essentially the market’s two-sided bet on which of those it is.
Evidence
Analysis
June 12 isn't just a SpaceX event. It's the first public price discovery for frontier AI equity. Every bank advising Anthropic and OpenAI will read the SPCX open and close before fixing their clients' terms. The second-order effect on the Q4 IPO window is the real number to watch.
What Investors Should Watch Before and After June 12
Between now and the June 12 open, three things matter.
First: confirmation or clarification on the Morningstar $780 billion estimate. If Morningstar issues a public note, the institutional market has a named reference point for the discount case. If the report remains unconfirmed, its influence is real but its precision is uncertain.
Second: the AFT SEC petition outcome. A formal SEC response to the pension fund scrutiny request would be the first regulatory friction signal in the offering and could affect institutional allocation decisions.
Third: roadshow demand signals. Pre-IPO books are typically oversubscribed for offerings of this size, what matters is the quality of the demand, meaning whether institutional buyers are committing at the $135 price or pressing for a lower allocation price.
After June 12, the first-day trading price is the data point. A close above $135 validates the IPO price and signals public market appetite for frontier AI equity at premium multiples. A close below $135, especially a substantial discount, tells every bank advising Anthropic and OpenAI that the public market’s pricing tolerance for pre-profitability AI infrastructure companies sits meaningfully below the private market’s recent valuations.
The real story is this: June 12 isn’t just SpaceX’s IPO. It’s the first real price discovery event for the entire frontier AI sector. Watch the open. Watch the close. The second-order effect on Anthropic’s October window is the number that matters most for investors who don’t own SPCX but do own, or will own, frontier AI equity.