In 2010, around the time of Tesla’s Initial Public Offering (IPO), an acquaintance proudly showed me his red Tesla Roadster, the first imported into Australia.
It was a fine little rocket, the first electric vehicle of the modern era, discontinued in 2012 having provided the engineering and financial foundations for the Model S and, eventually, SpaceX.
The company had sold only 1,000 of them in 2010 and was losing money but it floated for $US1.7 billion and then popped 40 per cent on its first day of trading, the first sign of Elon Musk’s pied piper power over investors.
The latest, and ultimate, sign of that came on Friday with the stock exchange debut of SpaceX.
The company had IPOd a few days earlier for $US1.77 trillion, almost exactly 1,000 times Tesla, and ended up trading 20 per cent higher on day one for a valuation of $US2.1 trillion ($3 trillion) — about the same as the total value of all companies on the ASX or all houses and apartments in Victoria.
Elon Musk is the first trillionaire, knocking off King Croesus of what’s now Turkey (550BC) and the previous richest person in history, Mansa Musa of Mali (1,300AD).
To get a sense of the difference between the Tesla and SpaceX IPOs, consider that 1.7 billion seconds ago was 1972, when Elton John released the prescient Rocket Man, while 1.7 trillion seconds ago was 52,000BC, when we shared the planet with Neanderthals.
Value is irrelevant
So is SpaceX worth $3 trillion?
Of course not, nowhere near it. The company is basically a good internet company with an unprofitable rocket business and a substandard AI chatbot, and it loses money.
But valuation metrics are irrelevant here. No-one is valuing it in any normal way, which is a testament to the incentives and persuasive powers of Elon Musk and Wall Street banks: 19 of them, basically the entire US investment banking industry, shared US$500 million in fees, with Goldman Sachs and Morgan Stanley getting US$100 million each.
The Financial Review’s James Thomson called it “not so much a company as an idea”, while economist Paul Krugman described Elon Musk less flatteringly as “a human Ponzi scheme”.
But maybe there’s another, simpler, name for what’s going on here: venture capital.
The willingness of optimists to hand over money for someone else’s dream is what fires progress in capitalism.
It’s always high risk and usually fails, so it’s done away from the public eye with private presentations and only goes public when it goes well and the dreamers get their money back through an exit: an IPO or a sale.
But most of the time it doesn’t work, the dream is a bust and no-one ever speaks of it again.
Elon Musk’s particular genius is to inspire millions to do it at once, in public, partly because Tesla worked out so well for everyone and partly because his dreams are so big, and well, interesting: batteries and EVs, brain-computer interface (Neuralink), tunnelling (The Boring Company), satellites, space travel, artificial intelligence and robots.
(He is also a “monstrous human being” according to Edward Luce, writing in the Financial Times over the weekend about Musk’s use of X, formerly Twitter, to promote racial hatred in the UK among other incendiary posts and reposts.)
The gap growing bigger
Apart from Musk himself, SpaceX’s creation of thousands of new millionaires and billionaires is a significant development in global inequality.
A week before the SpaceX IPO, Thomas Piketty’s Global Justice Project proposed a global wealth tax, with the money to go into a “global justice fund” to deal with inequality, although they didn’t specify what percentage it should be, and there’s absolutely no prospect of such a thing occurring.
An earlier idea from something called the International Tax Observatory suggested an annual tax of 2 per cent on billionaires’ wealth, which would cost Elon Musk $30 billion a year, something he would no doubt devote quite a lot of X posts to fight.
In any case, despite its name, SpaceX is mainly an AI event, adding to the roughly $20 trillion in new wealth created in the market values of the small group of companies involved in AI since the launch of ChatGPT in November 2022.
On page 11 of the SpaceX prospectus is a chart displaying what the promoters say is the “total addressable market” (TAM) for its products. It’s a total of $US28.5 trillion, of which $US26.5 trillion, or 93 per cent, is AI. By the way, the total is a ludicrous fifth of global GDP.
For their $US100 million fee, analysts working for the lead underwriter, Goldman Sachs, forecast that SpaceX’s total revenues would reach $US474 billion by 2030, up from $US18.7 billion in 2025, most of which would be driven by a 100-fold increase in data centre revenue in four years, from $US3.2 billion to $US322 billion.
They don’t specify how much of that would be data centres in space but that is clearly going to be SpaceX’s main business model.
Loading
Data in orbit
Orbital data centres are theoretically feasible and are being proposed because the solar panels making the electricity would be closer to the Sun and operate 24 hours a day, apart from eclipses, and because communities on planet Earth are starting to push back at the vast buildings being erected near them that consume all of their power and water.
The way it would work (theoretically) is that clusters of satellites or “compute platforms” in low Earth orbit would carry racks of radiation‑hardened processors and storage, powered by large solar arrays and batteries.
Data would be sent to them either directly from ground stations using laser or a radio frequency uplink or from other satellites.
The orbital centres would operate AI training and inference in the same way they do on Earth, except it would have to be mainly for applications that can tolerate extra latency since it would take longer for the data to get back.
As for sending the answers back to the ground, the idea is that the system would only transmit compressed outputs and usually only the differences: what most proposals call “compute‑in‑orbit, ship the deltas”.
While the physics works, the economics doesn’t — yet. As a blog post from Spanish engineering and technology firm Sener put it: “Technically viable and strategically attractive, their mass deployment will depend on key advances in launch systems, thermal management and space-grade electronics.”
Specifically, most analysts say satellite launch costs would have to be about a tenth of what they are now. That’s possible but challenging in four years’ time to say the least and would presumably also squeeze SpaceX’s profit margins.
But the truth is that very few investors who bought in the IPO and since are thinking about such dreary detail and probably don’t expect to be still holding the stock in 2030 anyway.
They are just getting aboard the Elon Express, hoping to make 300 times their money like those who bought into the Tesla IPO eventually did.
Mind you, very few of them, if any, would still be holding it 16 years later, and SpaceX also looks more like “pump and dump” than “buy and hold”.
Alan Kohler is a finance presenter and columnist on ABC News. He hosts the podcast That’s Business with Alan Kohler in the ABC Business Daily feed on Friday. He also writes for Intelligent Investor.
