
01 SpaceX’s Nasdaq debut
On 12 June 2026, SpaceX pulled off the biggest IPO in history.
The shares were priced at US$135. They opened at US$150, surged as high as US$176 during the session, and closed at US$160.95, up 19.22% on the day. After hours, they kept climbing, reaching US$167.57.
By the close, SpaceX was worth roughly US$2.1 trillion. Overnight, it became the sixth-largest company in America by market value. The only companies ahead of it were Nvidia, Microsoft, Apple, Amazon and Alphabet — the absolute giants of global technology. SpaceX walked straight into that club.
The IPO raised US$75 billion, smashing Saudi Aramco’s 2019 record of US$29.4 billion and easily beating Alibaba’s US$25 billion US listing in 2014. More than 500 million shares changed hands on the first day, with trading value of around US$80 billion — more than four times Nvidia’s volume that day.
The deal also minted the world’s first trillionaire.
Elon Musk’s net worth crossed US$1 trillion. His SpaceX stake alone was worth about US$867 billion.
But open SpaceX’s financial statements, and the story looks very different.
Remember: the five companies ahead of SpaceX all make money.
SpaceX does not.
In 2025, SpaceX generated US$18.67 billion in revenue, up 33% year on year. Sounds decent? The problem is that it also recorded a net loss of US$4.94 billion.
The first quarter of 2026 was even more worrying. Revenue was US$4.69 billion, with growth slowing to 15%. Net loss was US$4.28 billion. In one quarter, SpaceX lost almost as much money as it did in the whole of 2025.
Since its founding in 2002, SpaceX has accumulated losses of roughly US$41.3 billion.
Break the business down into three parts and the picture becomes clearer.
The space launch business made US$4.086 billion in revenue in 2025 and lost US$657 million.
Starlink is the only profitable engine: US$11.387 billion in revenue and US$4.423 billion in profit.
The AI business generated US$3.201 billion in revenue, but lost a lot of money.
Even Starlink, the only business currently producing real cash, has a problem: ARPU is falling. Average revenue per user dropped from US$99 in 2023 to US$81 in 2025, and then to US$66 in the first quarter of 2026.
In plain English: Starlink is growing by cutting prices. More users, yes. But each user is paying less.
So here is the picture:
US$18.6 billion in revenue. US$4.9 billion in losses. US$2.1 trillion in market value.
Is US$2.1 trillion too expensive?
02 Who is paying for this US$2.1 trillion valuation?
Let’s compare.
Nvidia’s latest annual revenue was about US$215.9 billion.
Amazon’s was around US$716.9 billion.
Alphabet’s was around US$402.8 billion.
These are the companies SpaceX now sits beside in the market-cap rankings. They make hundreds of billions in revenue a year. More importantly, they make profits.
SpaceX made US$18.67 billion in 2025 and lost US$4.94 billion.
The others are printing money. SpaceX is still burning it. Yet the market is valuing SpaceX as if it belongs on the same shelf.
Now compare it with Broadcom, which is closest in market value.
Broadcom is worth roughly US$1.76 trillion. In 2025, it generated US$63.9 billion in revenue and about US$23.1 billion in net profit.
SpaceX is worth more than Broadcom, but has less than a third of Broadcom’s revenue. Broadcom makes US$23.1 billion. SpaceX loses US$4.9 billion.
One earns US$23 billion. The other loses US$5 billion.
And the lossmaker is valued higher.
So why can a loss-making company stand next to Nvidia and Apple?
Because Wall Street says it can.
Goldman Sachs predicts SpaceX’s AI revenue will grow by 9,900% in four years, reaching US$322 billion by 2030.
Morgan Stanley goes even bigger: it forecasts US$3.4 trillion in revenue by 2040. That would require revenue to expand 210 times in 15 years — a compound annual growth rate of 43.3%.
What does 210 times growth in 15 years mean?
In the history of the US stock market, only three companies have ever done it. But here is the important part: when they started, their revenue bases were all below US$100 million.
SpaceX is not starting from US$100 million. It is starting from around US$15.5 billion, with growth of only 18% that year.
Going from US$100 million to US$20 billion is one thing.
Going from US$15 billion to US$3.4 trillion is something else entirely.
Even Nvidia has not achieved that kind of ultra-long, ultra-high growth path.
So let’s be honest: the US$2.1 trillion valuation is not based on today’s business.
It is based on tomorrow’s story.
And that leads to the next question:
Who is paying for tomorrow’s story?
The answer is: retail investors.
In a normal IPO, retail allocation is usually around 5% to 10%.
SpaceX pushed it to 30%.
Fidelity lowered its normal IPO participation threshold from US$100,000 to US$500,000 all the way down to US$2,000.
Robinhood and other platforms had almost no meaningful threshold at all.
At first glance, this looks like Musk being kind to small investors.
But think about it for one second.
The lower the threshold, the larger the retail base. The larger the retail base, the hotter the subscription. The hotter the subscription, the stronger the IPO hype.
And that is not even the whole story.
US IPOs often come with anti-flipping rules. If retail investors sell too soon after receiving an allocation, it can affect their ability to participate in future IPOs, or reduce their allocation priority. The exact period differs by broker: some are 15 days, some are 30 days.
What does that mean in practice?
If you get allocated shares, you are encouraged not to sell in the short term.
And if you do not sell, you stay in the market and help absorb selling pressure.
Retail investors were not invited to share the cake.
They were invited to carry the plate.
This is the uncomfortable truth: for large IPOs, the initial valuation often already prices in a lot of future growth. Retail investors are far more likely to become exit liquidity than to make easy money.
And in SpaceX’s case, the company is not even profitable yet.
Musk also wants SpaceX included in the S&P 500 and Nasdaq-100 as soon as possible. If that happens, large amounts of passive money would automatically buy the stock, helping support the market value.
But the S&P 500 rejected the idea outright. It requires at least 12 months of listing history, positive net profit in the latest quarter, and positive combined net profit over the latest four quarters. It made clear that a high market cap alone does not earn an exemption.
SpaceX is still losing money. So it cannot get in quickly.
Nasdaq is more flexible. It introduced a fast-entry mechanism in May this year, so SpaceX may have a chance there. But the likely index weight would be tiny. The passive inflow would be a drop in the ocean compared with a US$2.1 trillion valuation.
In the short term, the index route cannot save the story.
03 A feast where everyone gets what they need
If retail investors are paying, who is collecting?
Start with Wall Street.
SpaceX’s core underwriters include Goldman Sachs, Bank of America, Citigroup, JPMorgan and Morgan Stanley.
These banks are not just underwriters. They are also lenders.
Back in 2022, when Musk acquired Twitter, Morgan Stanley led a group of seven banks that lent about US$12.5 billion. After Twitter’s advertising revenue collapsed, those loans were stuck on the banks’ books. They wanted to offload them, but could not.
Then xAI was founded and borrowed about US$5 billion for AI infrastructure.
In 2025, xAI acquired X. Old debt and new debt together added up to roughly US$17.5 billion, all sitting under xAI.
Then xAI was folded into SpaceX, and the debt came along with it.
So how do you repay it?
You go public.
Before the listing, SpaceX borrowed another US$20 billion in bridge loans from the same group of banks. New money was used to deal with old money. At the same time, the bridge loan became the entry ticket for underwriting the IPO.
Morgan Stanley, in particular, had a very good deal. The Twitter loan that had been trapped for three years could finally be worked out through the IPO — and the bank could earn underwriting fees on top.
Lend money to the company.
Help it sell shares.
Then publish forecasts that support the valuation.
Goldman says four years, 9,900% growth.
Morgan Stanley says 15 years, 210 times growth.
Are these numbers independent research?
Or are they supporting documents for a very important client?
Interesting.
Now look at Musk.
SpaceX had been private for more than two decades. Musk had said many times that he was in no hurry to list it.
So why the sudden change?
Because he can no longer afford not to.
Musk’s Tesla compensation package will not be fully unlocked until 2035.
The debt from the Twitter acquisition has not fully gone away.
AI spending is becoming more aggressive by the year.
In 2023, SpaceX had almost no financing debt. By 2024, that number had risen to US$11.8 billion. By 2025, it had surged to US$26.35 billion.
Musk says he wants to go to Mars by the time he is 80.
He cannot wait.
The more important point is this: the money raised from the IPO is not mainly going into rockets.
In 2024, SpaceX’s equity investment was only US$6.5 billion — less than a single OpenAI funding round.
In the first quarter of 2026, SpaceX spent 76% of its capital expenditure on AI data centres. Rockets got only 10%.
So the company is rushing to list, rushing to raise money, and rushing to burn cash on AI.
Musk has gone from being one of AI’s loudest sceptics to using the AI story to push up valuation.
That is where he is now.
04 Do you believe in Mars?
The person rushing to list the company — how much control does he keep?
SpaceX uses a dual-class share structure.
Class A shares carry one vote each.
Class B shares carry ten votes each.
Before the IPO, Musk controlled 85.1% of the voting power. After the IPO, that only fell to 84.4%.
Class B shareholders have the right to elect more than half of the board. No matter how many future funding rounds happen, Musk can still control the board.
Even America’s three largest public pension funds issued a joint warning, calling the structure “novel and extreme”.
There is another detail in the incentive plan.
No matter whether the performance targets are eventually achieved, the voting rights attached to the awarded shares take effect immediately from the grant date.
In plain English: the voting power is handed over first.
Then there is the issue price.
US$135. Fixed pricing. Take it or leave it.
Normally, IPO pricing is the result of a negotiation with institutional investors. Musk effectively took that pricing power back.
Why could he do that?
Because demand was more than US$300 billion. The deal was nearly four times oversubscribed. If you do not buy, someone else will.
SpaceX raised US$75 billion while diluting only 4.2% of its shares.
When Alibaba raised US$25 billion, it diluted about 15%.
Open SpaceX’s prospectus and you will find things that would be unimaginable in the prospectus of a normal company.
The mission statement says SpaceX aims to “build the systems and technologies necessary to make life multiplanetary, understand the true nature of the universe, and extend the light of consciousness to the stars.”
This is not a prospectus.
This is science fiction.
Musk’s equity incentive package includes one billion restricted shares.
There are two unlocking conditions.
First, SpaceX’s market value must rise from US$500 billion step by step to US$7.5 trillion.
Second, SpaceX must establish a permanent human settlement on Mars with at least one million residents.
Both conditions must be met.
One million people living on Mars has been written into a CEO compensation agreement.
I have seen companies sell dreams before.
I have not often seen dreams sold like this.
And the Mars story keeps going.
Tesla’s Optimus robots as the advance team.
Glass-domed cities.
Terraforming so that humans no longer need spacesuits.
Every two-year launch window sending 1,000 to 2,000 spacecraft back and forth.
The Moon as a transfer station.
The AI story is also full of ambition.
SpaceX wants to deploy an orbital AI compute satellite constellation as early as 2028.
Its AI1 satellite would have a deployed wingspan of about 70 metres, larger than a Boeing 747-8.
It has applied to regulators to launch up to one million space-based data-centre satellites.
SpaceX describes itself as the only company with a commercially viable path to building orbital AI compute systems at scale.
Can Musk eventually deliver on these promises?
SpaceX has existed for more than twenty years and is still loss-making. Its accumulated losses are already above US$40 billion.
Mars colonisation?
Humanity has not yet sent a single person to Mars.
One million people living on Mars?
There is not even a shadow of that yet.
Orbital AI satellite constellations?
Still at the PowerPoint stage.
The US$2.1 trillion valuation is all being placed on tomorrow.
They believe tomorrow will arrive.
Do you?

