Nobody knows what SpaceX is worth

The position, and the loss

SpaceX is the other nine percent of my account, the part that is not Tesla. I bought it at the IPO at $135. It closed at $108.37 on July 31, 2026. I am down about 20% on it.

I am writing this while losing money on the position, which is the only honest time to write about one. It is also a strange moment to publish, because next week the company reports earnings for the first time in its life and then close to a billion shares stop being locked up. Whatever this note says, the market is about to say something louder.

The stock listed on June 12, 2026 at $135, closed its first day at $160.95 and peaked at $225.64 four days later.[X-02] From that high it is down 52%.

Four percent of a company is not a price

SpaceX has about 13.2 billion shares outstanding. Roughly 550 million of them were sold at the IPO.[X-02] That is about 4% of the company doing all of the trading, in what was otherwise the largest listing ever recorded.

That single fact explains the round trip better than any story about rockets. When Nasdaq fast-tracked the stock into the Nasdaq-100 in early July, index funds had to buy, and they had to buy from a float that small. Front-running that flow is what made the June high. The fade afterwards was not the market changing its mind about SpaceX. It was the flow ending.

A four percent float does not discover a price. It amplifies whoever shows up. I think the June number was a supply artifact rather than a valuation, and I think the same thing is true in reverse right now, which is the only reason I am comfortable being down and still buying.

You are not buying a rocket company

This is the part I think most people holding SPCX have not fully absorbed. On February 2, 2026, SpaceX absorbed xAI in an all-stock merger that valued the combined company at $1.25 trillion.[X-09] Then on June 16, four days after listing and on the exact day the stock peaked, it filed an agreement to buy Cursor, the AI coding company, for $60 billion in stock.[X-08]

So the ticker is rockets, plus Starlink, plus a frontier AI lab, plus an AI coding company. Two of those three were bought with shares, not cash.

Here is the detail almost nobody quoting the $60 billion headline mentions. The Cursor exchange ratio is not fixed in shares. It is set by SpaceX’s own volume-weighted average price over the seven trading days before the deal closes, against a fixed $60 billion of value.[X-07] Read that again with the chart above in mind: the lower SPCX trades into the close, the more shares existing holders give up for the same company. The deal is expected to close this quarter. A falling stock price is quietly making its own acquisition more expensive in dilution, and that is a feedback loop, not a one-time cost.

Two valuations, four times apart

Morningstar puts fair value at $62 a share and has not moved it since the IPO.[X-04] Their case is not complicated: the company has no earnings, it is currently losing money, and it trades around 78 times sales against roughly 19 for Nvidia.[X-05]

The sell-side could not disagree more. Consensus sits near $231, 27 of 28 analysts rate it a buy, and Morgan Stanley is holding a $300 target into the print.[X-06]

That is a 3.7x disagreement about the same company in the same week. It is not a modelling quibble, it is two incompatible descriptions of what this business is. One of them is wrong by more than a trillion dollars.

One thing worth saying in Morningstar’s favour, since the gap has moved: at the June high the stock traded at 3.2 times their fair value, which made it the second most expensive stock in their entire coverage universe.[X-04] At $108.37 it trades at 1.75 times. The bear case is being paid attention to. That is a better reason to buy than any of the bull notes.

The week ahead

Two dates, back to back, and they are the reason this note is dated so precisely.

• August 4, 2026. The first public quarterly report in the company’s history.[X-01] Nobody outside SpaceX has ever seen real audited quarterly numbers for Starlink. Every model above, bullish and bearish, is built on estimates that are about to meet an actual income statement.
• August 6, 2026. 911.5 million shares come unlocked, with a further 455.8 million possible under certain conditions.[X-01] Musk and a group of insiders stay locked until the middle of 2027.[X-03]

Put the unlock next to the float and the size of it becomes obvious. The shares coming free are about 1.7 times the entire float that exists today, and the tradable share of the company roughly triples to about 12%.[X-02] At July 31, 2026’s price that is on the order of $99 billion of newly sellable stock. Published figures put it higher, between about $116 and $123 billion, because they were written when the stock was more expensive; the share count is the number that has not moved, so I did the multiplication myself.

Eligible to sell is not the same as sold, and I would bet most of it does not move. But this is the first time SPCX will have a real supply of shares meeting real demand. Everything before Thursday was four percent of a company trading on flow. Whatever price comes out the other side is the first one that means very much.

And now every launch is an earnings event

On July 24 Starship Flight 13 deployed all twenty of the first Starlink V3 satellites and the ship came down in the Indian Ocean in what SpaceX called its softest splashdown yet. The booster was lost: only ten of its thirteen engines relit, and five of those failed seconds later.[X-10]

Primary objective met, hardware destroyed. As engineering that is a good day, and V3 Starlink is the thing that actually pays for all of this. As a shareholder it is a new and genuinely uncomfortable fact: the company’s R&D now happens in public, on video, and the stock trades on it. Nobody who bought “the SpaceX IPO” signed up to have a quarter re-rated by whether a booster relights.

What I am doing, and what would prove me wrong

Small regular buys, and I am not adding size into the unlock. The whole reason this is nine percent of the account rather than ninety-one is that I cannot defend a price here. I can defend the company. Those are different claims, and mixing them up is how people end up concentrated in something they only half understand.

Written down now, so I cannot move them later:

• Starlink subscriber growth comes in flat or decelerating in the first two public quarters.
• The unlock produces sustained insider selling over months rather than a single bad week.
• Cursor closes on a materially lower average price and the dilution lands harder than the $60B headline implied.
• Two years from now Morningstar's $62 has proven closer than the sell-side's $231.
• Starship keeps losing boosters and the reuse economics never bend the cost curve.

If three of those are true by the end of 2027, I was wrong about this one, and the nine percent was the correct size to be wrong at.

The disclosure

None of this is advice. I am seventeen, it is a Fidelity Youth Account, I place the trades myself, and I own the stock I am arguing about while sitting on a loss in it. Read the bear case twice.

Sources

1. [X-01]First earnings August 4 and the August 6 lock-up: 911.5M shares, plus a conditional 455.8M
2. [X-02]13.2B shares outstanding, ~550M sold at IPO, and the float roughly doubling to about 12%
3. [X-03]Nearly a billion SpaceX shares unlock on August 6
4. [X-04]Morningstar's $62 fair value against the market price: “one of them is wrong by $1.2 trillion”
5. [X-05]Morningstar on why it thinks the SpaceX IPO is overvalued
6. [X-06]Morgan Stanley reiterates its $300 target going into the first report
7. [X-07]SpaceX formalizes the $60B all-stock merger agreement to acquire Cursor
8. [X-08]SpaceX to acquire Cursor for $60B in stock, days after the IPO
9. [X-09]The SpaceX and xAI all-stock merger, valuing the combined company at $1.25T
10. [X-10]Starship Flight 13: all 20 Starlink V3 satellites deployed, soft splashdown, booster lost
11. [X-11]SPCX price and the lock-up expiry in early August

WRITTEN 2026-07-31 · PRICES FROZEN AT JULY 31, 2026, THE LAST CLOSE BEFORE PUBLICATION. THIS NOTE IS DELIBERATELY NOT UPDATED AFTER THE EVENT.