SpaceX IPO Valuation Halves, Morgan Stanley Sees It Doubling Anyway

SpaceX completed its blockbuster IPO in June at a $1.77 trillion valuation. The stock proceeded to lose roughly a third of that value. Morgan Stanley responded by maintaining a buy rating and a $300 price target implying more than 100% upside. This is not a typo. This is what passes for market efficiency in 2024.

The numbers are real. SpaceX raised more than $85 billion at that $1.77 trillion initial valuation, then watched its market cap crest just under $2.8 trillion in the immediate aftermath before collapsing 36% from that peak. Current valuation sits just below $2 trillion. That's a $1.8 trillion company that lost roughly $800 billion in value faster than most people can update their portfolio apps. Most investors, presented with this sequence of events, ask one question: what went wrong?

Morgan Stanley asks a different question: what's the AI opportunity?

According to the firm's September analysis, $26.5 trillion of the company's total $28.5 trillion addressable market comes from AI operations. This is the architecture of the bull case. SpaceX's valuation, in other words, does not rest primarily on rockets or satellites or launch services or even Starlink's satellite connectivity network. It rests on the proposition that SpaceX will capture meaningful economic value from artificial intelligence compute and infrastructure. The space business is, by this accounting, scenery.

This creates a problem that 20 years of watching traders should have taught Wall Street to notice. The company's AI division currently operates at a significant loss. In the most recent quarter, SpaceX deployed $18.37 billion in capital expenditure, with $15.83 billion dedicated to AI compute infrastructure. Revenue jumped 92% to $7.8 billion. Adjusted EBITDA climbed 191%. The company is spending its way deeper into red ink at the largest scale in corporate history in pursuit of a market that doesn't yet exist at commercial scale.

Morgan Stanley's thesis requires that this changes. The $300 price target requires that AI operations mature from capital sink to profit center at a velocity that justifies a 100% revaluation of a company that already trades at nearly $2 trillion. A $100 share price, Morgan Stanley's own math suggests, would imply zero economic value for the AI division. Zero. The entire bull case compresses into a single variable: belief that SpaceX's AI infrastructure will eventually generate enough economic return to justify its $15.83 billion quarterly burn rate.

The market is apparently split on this. Twenty-eight Wall Street analysts cover the stock. Six rate it strong buy. Twenty-two rate it buy. Five rate it hold. Two rate it sell. Zero rate it strong sell. This is not skepticism. This is organized credulity.

What makes the Morgan Stanley view particularly instructive is not that it's bullish. It's that the bullishness is completely decoupled from the IPO experience. An investor who bought at the $1.77 trillion IPO valuation watched their position decline 32%. Morgan Stanley's response is not to ask whether the original thesis was flawed. It's to ask whether AI value captures have grown wider. The IPO was not a market test. It was a down payment on a belief system.

There is nothing inherently wrong with belief systems in equity markets. Markets function because people believe future cash flows matter. The problem arrives when the belief system becomes invisible. When the $300 price target doesn't rest on near-term revenue or EBITDA or cash generation, but instead on the resolution of a multi-year AI infrastructure bet that currently bleeds cash, it should require explicit acknowledgment that you are no longer investing in SpaceX. You are investing in a hypothesis about artificial intelligence infrastructure economics. The vehicle happens to be a space company.

The IPO crash was supposed to be the market's truth machine. Instead it was a volume adjustment. SpaceX is still valued at nearly $2 trillion. Morgan Stanley still sees it doubling. And the company will still spend $15 billion next quarter on AI compute infrastructure that has not yet generated a dollar of meaningful commercial return.

That's the math problem. Wall Street isn't troubled by it because Wall Street has redefined what it's solving for.