Pricing the Founder

The market is being asked to accept a $1.77 trillion price built less on what SpaceX earns than on what its founder promises it will become. There is a disciplined way to value that promise, and a reason every risk committee should care how this one is priced.

The number, and the problem with the number

At a fixed $135 per share, SpaceX would list at roughly $1.77 trillion, large enough to rank around the seventh-most-valuable company in the United States, ahead of Tesla. The structure is unusual in ways that matter to a diligence process. The price is fixed rather than discovered through a book-build. The initial float is around 5%, which mechanically amplifies the first move in either direction. Reporting suggests up to 30% of shares may be steered to retail, against a typical 5 to 10%.

The valuation problem is straightforward to state. The disclosed business, a consolidated SpaceX and xAI entity reporting roughly $18.67 billion of 2025 revenue and a $4.94 billion net loss, does not on any conventional multiple support a $1.77 trillion enterprise value. The gap is bridged by narrative. Starship, a million people on Mars, orbital data centres, and a founder whose track record invites the benefit of the doubt. Morningstar's published fair value sits near $780 billion, less than half the ask, and notes that only Starlink is currently profitable while xAI is projected to burn about $10 billion in 2026.

The disclosed business does not support the price. The gap is narrative, and narrative can be valued, but only if you are honest that that is what you are valuing.

A prudent methodology: sum-of-the-parts with a real-options overlay

Our position is that the responsible way to value SpaceX is to separate what exists from what is promised, value each on its own terms, and refuse to let optionality masquerade as base-case enterprise value. We propose three layers.

Layer one: the cash engine

Starlink is the load-bearing asset. Recurring, growing, and the one profitable line. It should be valued on satellite-telecom and recurring-revenue comparables, with explicit haircuts for churn, capex intensity, and competitive entry. The launch business is valued separately on contracted backlog and demonstrated cadence. This layer is conservative by design and produces a defensible floor.

Layer two: strategic optionality

xAI, Mars, Starship, and orbital compute are real options. Each has a credible but unproven path to value. The honest treatment is a probability-weighted net present value. A scenario tree that assigns a defensible chance to each outcome, computes a present value conditional on success, and weights accordingly. The output is not a forecast. It is a price for the option, separately disclosed.

Layer three: the narrative premium

Whatever remains between the sum of layers one and two and the ask is narrative premium. Call it what it is. Investors who pay it are buying access to a story they believe will be re-told at a higher price. There are conditions under which that is rational. Founder concentration, network effects, and option value all create reasons to pay above modelled value. But the premium should be sized, disclosed, and stress-tested. Not absorbed silently into the base case.

The governance discount

The compensation award disclosed in the S-1 has, on the lower bound, a current value near $175 billion, with a theoretical maximum upside near $1.1 trillion across twelve milestone tranches up to a $7.5 trillion market capitalisation. Contrary to a common framing, the award is not a thirty-year package. None of the bonuses is tied to a time frame, and the award persists for as long as the founder is employed. More consequentially, the restricted shares confer super-voting rights immediately upon grant, before any milestone is met. The practical effect is durable founder control that does not depend on Mars, Starship, or any milestone materialising.

A compensation consultant quoted on the filing described the milestone spectacle bluntly as marketing built to drive the price and the raise. For a governance reviewer, the question is not whether the targets are achievable. It is what minority shareholders own if they never are. The honest answer is: economic exposure to a business they cannot direct, in a structure that already vests control elsewhere. That should attract a discount.

The precedent

Every listing that follows will study this one. A fixed-price take-it-or-leave-it raise of this scale, with a thin float, an oversized retail allocation, and a pay package that grants control before performance, sets terms that other founders will ask to mirror. Risk committees, allocator boards, and policy desks should treat the SpaceX IPO not as a space story but as a precedent. The architecture of how a company comes to market is becoming a governance question in its own right.

What history rhymes with

Three precedents are worth holding in mind. Saudi Aramco listed in December 2019 at a $1.7 trillion valuation on a float of just under 1.5%, with the price effectively set by the sovereign rather than discovered through a competitive book. The early aftermarket was supported, the float was thin, and the eventual price action depended on conditions far outside the company's control. The lesson is not that fixed-price mega-listings fail. It is that the price prints what the sponsor wants it to print on day one, and the market gets its vote later.

Alibaba's 2014 listing at a $231 billion valuation cleared because the underlying business was already cash-generative at the scale being asked. The narrative premium was modest relative to disclosed earnings. SpaceX is the inverse case. The narrative premium is doing most of the work, and the disclosed business loses money. That asymmetry does not make the deal wrong. It does mean the buyer is underwriting more story per dollar than Alibaba's were.

Facebook's 2012 listing is the cautionary tape. A high-profile retail-heavy debut, a software-execution stumble at the open, and a six-month drawdown of more than 50% before the business caught up to the price. The business did catch up. Investors who held through were rewarded. The ones who sized too large were not in a position to hold. That is the pattern that matters for SpaceX. The question for any prudent buyer is not whether the company survives. It is whether the position survives the path.

What an institutional model actually outputs

Using the three-layer framework with defensible inputs, a base case anchored on Starlink at 35 million subscribers in 2030, $96 ARPU, a 30% terminal margin, and a 6x sales exit; plus a contracted launch backlog at 18% margin; plus xAI on a probability-weighted ramp; plus a 25% optionality block for Starship, Mars, and orbital compute, lands in the $850 billion to $1.1 trillion range. The bull case, with optimistic Starlink penetration and a 60% probability of xAI reaching escape velocity by 2030, reaches roughly $1.5 trillion. To get to $1.77 trillion requires either a narrative premium of 15 to 25% on top of an already-bullish stack, or a base case that assumes outcomes most underwriters would not put in writing.

That gap is the founder premium. It is not unreasonable to pay it. It is unreasonable to pay it without naming it.

The governance discount, quantified

Empirical work on dual-class structures by Bebchuk and Kastiel, among others, finds a persistent discount of 8 to 15% on companies where founder voting rights remain decoupled from economic interest for extended periods. SpaceX's structure is more concentrated than the median dual-class case studied. A 10% governance discount applied to a $1 trillion fundamentals stack removes $100 billion from defensible value before the narrative layer is even considered. Applied to the $1.77 trillion ask, the implied governance haircut is closer to $175 billion. Either number is a real piece of the price, and neither shows up in a standard multiples comparison.

The regulatory precedent

The SEC's posture toward fixed-price listings has been permissive in recent years, particularly where the issuer can demonstrate sufficient institutional demand at the offered price. That permissiveness assumes price discovery is happening somewhere in the process, even if not in the prospectus. With a 30% retail allocation and a book that may be filled before retail orders are even logged, the discovery happens in the secondary market, on the backs of buyers who cannot price-set.

Other regulators are watching. The FCA's listing-rules consultation explicitly cites super-voting structures and concentrated control as live questions. The Hong Kong Stock Exchange permits weighted voting only with sunset clauses. The European Securities and Markets Authority has flagged founder-pay structures with no time cap as a disclosure concern. If SpaceX prints clean on Friday and trades well, the architecture becomes the template. If it prints clean and trades badly, the architecture becomes the case study for tighter rules. Either way, the precedent is the durable output of this listing, not the share price.

What a risk committee should say in writing

For institutional allocators, the minutes of the IPO-week meeting should record four things. First, the explicit fundamentals value, computed from the floor and disclosed alongside the narrative premium it implies. Second, the governance discount applied, with the methodology cited. Third, the position-sizing decision relative to liquid capital, with a stated maximum drawdown the allocation can absorb. Fourth, the rebalancing trigger, written down before the print, that says at what price the position is added to, trimmed, or exited.

The same discipline applies to a household balance sheet, scaled down. The point of writing it before Friday is to make the Friday decision smaller than it would otherwise be.

The physics of the milestones

Two of the twelve tranches in the founder package condition on outcomes that no spreadsheet can value alone. The first asks for roughly 100 terawatts of compute, explicitly including off-Earth capacity. The second pairs a $7.5 trillion market capitalisation with a self-sustaining settlement of at least one million people on Mars. Pricing them honestly means treating them as physics before treating them as options.

At hyperscale, the binding constraint on orbital compute is not power and not cost. It is thermal: in vacuum, heat leaves only by radiation, and Stefan–Boltzmann sets the radiator area required at roughly 1,200 square metres per megawatt of waste heat at electronics temperature. At 100 terawatts, the area is measured in millions of square metres per facility, before any rad-hardening derate is applied to compute-per-watt. On Mars, the binding constraint on a million-person settlement is not transport cost. It is the dose-versus-light trade: protecting people from 240 to 300 millisieverts per year of background radiation requires burying them under one to several metres of regolith, which makes natural-light access a rationed exception. The companion piece, The Physics of the Premium, runs both simulators end-to-end. The relevance to this article is that the probability an institutional model should assign to each milestone vesting on any plausible horizon is informed by the physics, not the narrative.

What we recommend

Build the valuation from the floor. Anchor on Starlink and the launch backlog. Add a sized, disclosed optionality layer. Refuse to call narrative premium anything other than what it is. Apply a governance discount that reflects the durable control conferred at grant. Compare the result to the ask, not to a peer multiple. If the gap is uncomfortable, that is information, not an instruction to buy.

The listing itself is the easy part. The next ten years are the part that decides whether the price made sense. Every framework above is built to survive that distance, not to win the open.