Three weeks ago, Elon Musk’s SpaceX was valued by the market at more than $2.5 trillion ($3.6 trillion). Its value fell below $2 trillion on its first day of inclusion in the Nasdaq index, which was expected to trigger a wave of forced index fund buying.
Tuesday’s 6.83% stock decline narrowed SpaceX’s decline to about 22% from its high on its second day of trading last month. Mandatory index fund purchases to reflect index fund weighting in the Nasdaq index have so far failed to materialize.
Even a series of ultra-bullish reports from six of Wall Street’s most prominent investment banks (Morgan Stanley’s “bullish” claim on the company, with a price target of $600 per share or a market cap of more than $8 trillion) didn’t help.
In the bond market, spreads on SpaceX’s $25 billion raise are blown out. Initially, the spread was set at 1.4 percentage points over U.S. Treasuries, but the spread widened to 1.65 percentage points. Investment-grade bonds are trading in junk bond territory.
SpaceX is not alone. The uncritical enthusiasm for all things related to mega-technology and artificial intelligence has waned.
The giant tech stocks known as the Magnificent Seven are down 6.5% from their recent highs in May, despite rebounding last month. The PHLX Semiconductor Sector Index has fallen 16% in two weeks after posting a massive 105% AI-led rally at the beginning of the year.
In the bond market, the $25 billion Amazon bond issue raised only about half the interest of its previous issue in March, while the prices of other AI-related bonds fell and their yields rose (prices and yields are inversely correlated), showing that the market was satisfied with AI companies’ debt.
The state of the equity and bond markets regarding AI issuance suggests a degree of fatigue and perhaps nervousness.
When SpaceX emerged, investors were primed to believe it would bring extraordinary value to things that exist only in Elon Musk’s head: data centers in space and the colonization of Mars, in an environment where the value of AI wannabes is only increasing.
OpenAI and Anthropic (both of which raised less than $200 billion last year) have been discussing and filing draft prospectuses for their own multitrillion-dollar floats in the wake of SpaceX’s spectacular debut.
OpenAI is now talking about floating, possibly next year, which will inevitably make Anthropic rethink its current market entry.
AI seems like a winner-take-all industry. Companies will either emerge as the dominant force in the AI space or disappear.
The sudden volatility in AI stock trading comes as concerns continue that the sector is in bubble territory after four years of AI-driven bull markets.
This is making some investors nervous. The insatiable appetite for capital to fund AI investments relies on continued access to stock and bond markets on ever-improving terms.
If that access were to be reduced or the price of access to be significantly increased, the entire sector could collapse, with unpleasant implications for broader markets and the economy, particularly the U.S. economy.
Pure AI companies like OpenAI, Anthropic, and even SpaceX, whose valuations are dominated by AI-related potential, must continue to raise equity and even debt to fund ever-increasing investments in computing and supporting infrastructure.
To attract that equity and debt, valuations based on earnings prospects for companies with mounting losses today must continue to rise. Otherwise, they will lose access to capital and will not be able to maintain the spending needed to keep up with more diversified competitors, leading to bankruptcy.
Even hyperscalers like Google, Meta, Amazon, and Microsoft have huge non-AI cash flows that limit their ability to fund the level of capital spending determined by their AI ambitions.
They plan to invest over $750 billion in AI this year, over $1 trillion next year, and even more after 2028, which will outstrip even their ability to generate cash. So even Amazon and Google’s parent company Alphabet are raising new capital and debt.
AI seems like a winner-take-all industry. Companies will either emerge as the dominant force in the AI space or disappear.
Until very recently, investors didn’t discriminate against companies fighting AI (with the possible exception of highly leveraged Oracle, whose stock price fell 43 percent in one month), but that may be changing as investors assess which of their competitors has the ability to survive to the end and ultimately emerge with super-profits that justify the extraordinary levels of investment.
Of course, it is possible that such super profits will not materialize.
To date, growth in AI investment has outpaced revenue growth, and the gap is widening as the size of capital investment gradually increases, and while AI adoption by companies has been significant, it has been slower than expected as AI has not yet delivered the productivity gains relative to cost that AI companies envisioned and that companies expected.
AI is also becoming increasingly commoditized, and the benefits gained from the launch of a new frontier model are erased within weeks by the release of a competing model. Cheap Chinese AI models are sucking up demand for less sophisticated AI capabilities.
If neither AI proponents nor their business and consumer customer bases are able to earn adequate returns on the capital they are risking, the sectors’ current business models of investing at exponential growth rates and generating exponential revenue growth will collapse.
Draft US Treasury report obtained by Digital News Agency Notus (News of United States), like many in the market, pointed to the similarities between AI and the dot-com boom, concluding that AI companies are more entrenched in the U.S. economy than dot-com companies and pose significant risks to the financial system and economy if financial conditions change, productivity goals are missed, or challenges impede growth.
The financial landscape may be changing, which could be a factor in recent AI deals. Challenges are emerging as data center developers and semiconductor manufacturers struggle to maintain the pace needed to meet demand.
The yield on the two-year Treasury note, which is most sensitive to current financial conditions, rose 13 basis points to 4.19% since last month’s Federal Reserve meeting raised the prospect of rate hikes this year.
The 10-year Treasury yield rose 11 basis points to 4.55%. The 10-year bond provides a “risk-free” interest rate that is the center of the discount rate used to calculate the net present value.
For companies that are priced for perfection, such as AI companies, dizzying valuations are predicated on huge revenues in the not-too-distant future and sustained growth well into the future, so if interest rates rise significantly, the notional value will decline.
If this does indeed lead to a decline in stock market valuations, it could cause a meltdown, given the increased financial leverage within the industry and the operational leverage created by the myriad contractual relationships between all of the major AI participants.
Hyperscalers with non-AI cash flows may survive, and indeed emerge as winners in the desperate scramble for AI supremacy, but key parts of the sector—AI developers, chip makers, data center operators, and others providing the “picks and shovels” for the AI boom—will suffer.
Until recently, investors were prepared to ignore the sector’s potential risks, valuing everything about AI as if there was a seamless, risk-free path to realizing its potential, including data centers in space and colonies on Mars.
That may be changing now, at least with greater awareness of the risks. Only experience and hindsight, perhaps cruelly, can tell whether that will lead to a drastic, bubble-like drop in valuations that could threaten the broader financial system and economy, or just a pause in the cycle.
Our Business Briefing newsletter brings you top stories, exclusive coverage and expert opinion. Sign up to receive it every weekday morning..
