Everything about the SpaceX float is bigger than Texas.
There are the sheer numbers, which were released by the cadre of bankers on Wall Street on Wednesday night: the company will raise at least $US75 billion ($105 billion) at a valuation of at least $US1.75 trillion ($2.45 trillion). Should the bankers really try to squeeze every bit of juice out of this deal, the valuation could rise to $US1.8 trillion.
This will be the biggest float in history. And even though the free-float – the amount of stock actually being sold off to new investors – is small compared with SpaceX’s valuation, it is still more than double the near $US30 billion that oil giant Saudi Aramco raised in its $US1.7 trillion initial public offering in 2019.
The total addressable market SpaceX claims is similarly gigantic, at a staggering $US29 trillion. This is not built on SpaceX’s actual space ventures – including its very real and quite profitable Starlink satellite business and its very unproven data-centres-in-space pipedream – but on selling the same artificial-intelligence services to the world that Anthropic and OpenAI are now desperately trying to flog to businesses globally.
Then there’s the raw power that SpaceX founder, chairman, chief executive and chief technology officer Elon Musk will hold over the company, courtesy of his 82 per cent voting stake.
And finally, there’s the risk. Layers and layers and layers of enormous risks.
Neither profitable nor predictable
SpaceX is not a highly profitable, highly predictable company like Saudi Aramco but a conglomerate that lost almost $US5 billion last year, and is likely to keep losing money for at least the next two years and possibly longer. And yet, it is about to float on the public market at an eye-watering valuation that is equivalent to about 6 per cent of US GDP.
Former Lehman Brothers trader Larry McDonald, the founder of research service The Bear Traps Report and author of bestseller How to Listen When Markets Speak, says that’s 847 times more expensive than the Microsoft IPO in 1986, 653 times more expensive than the Tesla IPO in 2010 and 14 times more expensive than the Facebook IPO in 2012.
That’s, of course, if you actually believe the $US1.75 trillion valuation. Analysts Nicolas Owens and Suryansh Sharma of Morningstar, a firm that famously relies on the Warren Buffett theory of economic moats to assess companies, could get their valuation of SpaceX to only $US780 billion. That is still pretty impressive, at more than double BHP’s market capitalisation.
The Morningstar team’s big problem was that the economic moat around SpaceX’s AI business – again, this is where SpaceX itself says the really big opportunity lies – is “indeterminate, and it also poses a material threat of value destruction to the company”. Gulp.
If you’re willing to cast aside those concerns about fundamental valuation and put your faith in the admittedly proven ability of Musk to create value by selling a vision for the future, then you need to ask another question: why now?
Why is it that SpaceX, Anthropic and OpenAI are suddenly in this urgent race to list on public markets at valuations that will probably make all of them trillion-dollar companies? Dario Perkins, global macro strategist at TSLombard, asks the question another way: “If AI is going to transform the world, why are tech giants deciding to share the wealth with normies? Perhaps they know something we don’t.”
Big-bang IPOs that marked the end of booms
Yes, access to capital markets will be vital, given the funding requirements of this AI revolution, which is only getting more expensive. But Perkins says history is replete with examples of big-bang IPOs that marked the end of booms.
There was the float of Blackstone in 2007, just before the global financial crisis. There was the IPO of Glencore in 2011, which marked the end of the mining super cycle. The miner has traded above its issue price for only one month of its life as a public company. And then there was a string of top-of-the-cycle IPOs in the late 1990s, including VA Linux, Pets.com and AskJeeves. With most of those deals, company insiders were prevented from selling their shares for periods of about six months – just as in the SpaceX IPO.
Those lock-ups expired from around October 1999 to April 2000. Within months, the dotcom bubble had burst. “Insiders know when public investors are overvaluing the business and decide to get out while the times are good,” Perkins says.
To be clear, all of these risks are in plain sight – Musk is not pulling the wool over anyone’s eyes here.
Markets have tweaked the rules
But hundreds of millions of investors around the world will be exposed to this risk whether they like it or not, thanks to recent decisions by index providers including FTSE Russell, Nasdaq and CRSP, which have tweaked their rules in the past few months to allow mega floats such as SpaceX a much faster entry into benchmark indices than in the past.
“The concentration risk that exists in equity markets hasn’t been this great in 100 years.”
Once upon a time – about three months ago – inclusion in these indices involved passing strict hurdles. A company would have to wait about three months after listing to be considered for inclusion, have a free float (shares not controlled by existing shareholders) of more than 10 per cent, and be profitable for a certain period.
But with the floats of SpaceX, Anthropic and OpenAI on the horizon, these rules have suddenly been changed. SpaceX will become eligible for inclusion in the Russell 2000 index within just five days and in the Nasdaq indices within 15 days. S&P Dow Jones, the world’s biggest index provider, is due to announce its new rules on June 8. Competitive forces will surely force it to fall in line.
But what’s already clear is that these rule changes mean the hundreds of trillions of dollars tracking indices through passive strategies – including the world’s most popular exchange-traded funds – are going to be required to buy shares in companies such as SpaceX almost as soon as it lists, in what’s certain to be a blaze of bubbly hype.
That includes tens of millions of Australians, both retail investors who hold ETFs directly and super fund members, whose funds are increasingly reliant on passive or passive-like strategies across their equity portfolios.
Some commentators, such as The Bear Traps Report‘s McDonald, say the index providers have failed to protect investors by changing their rules to allow fast entry into key benchmarks. They believe investors’ retirement savings are about to be hijacked and used as exit liquidity for insiders at companies such as SpaceX, as they sell shares over the next 12 months.
‘Dangerous, dangerous territory’
Anton Eser, the global chief investment officer of Dutch funds management giant Robeco, isn’t so worried about that. He points out that most benchmark indices are weighted according to the free-floats of their constituents, and the free-floats of SpaceX, Anthropic and OpenAI are likely to be small, at least initially.
But Eser, whose firm manages about $425 billion and about $65 billion of Australian institutions – most of which is held in quantitative strategies that broadly track sharemarket indices, and then find small pockets of above-market returns – says the float of SpaceX and other giants should force all investors to consider just how concentrated big sharemarket indices have become from the AI trade.
The best estimate, from Bank of America, suggests that when SpaceX, Anthropic and OpenAI list, about 48 per cent of the S&P 500 will be weighted to 13 AI-related companies. We haven’t seen anything like that since the railroad boom of the late 1800s.
“The concentration risk that exists in equity markets hasn’t been this great in 100 years,” Eser tells Chanticleer during a visit to Australia. “We are in dangerous, dangerous territory.”
What he particularly struggles with is modelling the profits that the companies selling AI products and services – such as SpaceX, Anthropic, OpenAI and other hyperscalers – can actually generate.
He’s not alone. Committed AI sceptics such as Ed Zitron have long argued that the return on investment on AI is illusory because the value it creates is illusory. But new concerns about rising AI costs have brought these concerns into the mainstream.
A new report from consulting giant Bain & Co this week found 40 per cent of companies that have measured cost savings through AI have cut costs by 10 per cent or less. “The technology worked. The value didn’t arrive,” Bain concluded.
If AI is not generating returns for users, how will it generate earnings for companies such as SpaceX and the other firms pouring trillions of dollars into AI infrastructure? If the faith of AI customers is eroded, will the brakes be applied to the AI buildout?
If nothing else, the SpaceX float reminds us that every passive investor and every super fund member is now incredibly exposed to a sector whose economic model is still largely built on hope.