SpaceX Stock Sits Just $15 From a Zero Value AI Verdict

SpaceX stock closed at $115.07 on July 24, just $15.07 above the $100 level Morgan Stanley says would price the company’s entire artificial intelligence business at nothing. That is a 28.5% drop from the $160.95 close on the Space Exploration Technologies Corp. (SpaceX) initial public offering (IPO), and it has erased roughly $1 trillion of the nearly $3 trillion peak value the stock touched on June 16, when shares hit $201.80.

The AI unit sitting at the center of that math holds Grok, the xAI chatbot; X, the social platform formerly called Twitter; and Cursor, the coding startup SpaceX folded in for $60 billion in stock barely six weeks ago. Wall Street’s own numbers currently price that whole package at zero, maybe less, even as the company’s rocket and satellite businesses keep printing real cash.

The Fifteen Dollar Line

Morgan Stanley analyst Adam Jonas, who covers the stock for the bank, laid out the math in a note to clients Friday, according to Bloomberg. He kept his Overweight rating, Wall Street’s label for a stock he expects to beat the market, and a $300 price target that implies a valuation near $4 trillion, a figure that splits opinion across six other banks covering the stock.

Most investors we speak with significantly discount Grok & Cursor. Many ascribe zero or even negative value for AI given the high capex requirements relative to Space & Connectivity, largely uncertain economics, and the high degree of management time devoted to the business.

Jonas wrote that line to explain why he thinks the sell-off has run ahead of the company’s actual results. He told clients that ‘the current disconnect between increasingly bearish investor sentiment and largely unchanged fundamentals creates an attractive entry point in SpaceX shares.’ Translation: he thinks the stock is being punished for sentiment, not for anything that broke in the business itself.

Three Businesses Share One Stock Price

SpaceX now reports results in three pieces: Connectivity, essentially Starlink; Space, the launch business; and AI. The first quarter of 2026 shows how lopsided those three have become.

Add the three together and SpaceX took in about $4.7 billion for the quarter against a combined operating loss near $1.9 billion. The company’s actual bottom line for the quarter was still a $4.27 billion profit, a gap the filings attribute to interest income, investment gains and other items that are not broken out by division.

Starlink Is Carrying Everyone Else

Connectivity is the reason SpaceX makes money at all. The segment brought in $11.39 billion in revenue during 2025 and turned an operating profit of $4.42 billion, up nearly 50% from a year earlier, in a segment breakdown of the IPO filing published by Morningstar.

That is almost the exact revenue Verizon booked in 2025, also $11.39 billion, but Verizon needed 129 million combined wireless, broadband and video connections to get there. Starlink did it with 10.3 million subscribers across 164 countries, a sign of how much more each Starlink customer pays for the service.

The business is not friction-free. Trade coverage and regulatory records point to a handful of recurring problems:

• Network congestion is building in some of Starlink’s most heavily used markets.
• Customer service has drawn close to 1,000 complaints filed with the Federal Communications Commission (FCC) over five years, with 36% of them mentioning support and 28% referencing open tickets that went nowhere, according to a Fast Company review of the agency’s records.
• The dishes need a clear view of the sky and cooperative terrain to perform well.
• Communities and regulators are pushing back on the pace of new satellite launches Starlink needs to keep expanding capacity.

That has not dented the segment’s profit growth so far, which is exactly why analysts keep pointing to Starlink as the part of SpaceX propping up the rest of the ledger.

What SpaceX Got For $60 Billion

The AI unit tells the opposite story. It lost $6.36 billion in 2025 on $3.2 billion of revenue, and the trend is worsening rather than improving: in the first quarter of 2026 alone it lost $2.47 billion on $818 million in revenue, a loss triple the size of its sales.

Some of that spending bought Cursor, the code-editor startup SpaceX absorbed into the AI unit in mid-June for $60 billion in stock. One early result is Grok 4.5, a coding-focused model trained using Cursor’s tools, priced at $2 per million input tokens for its entry tier. DeepSeek, the Chinese lab that keeps undercutting American labs on price, sells its cheapest tier for roughly $0.01 per million tokens, about a twentyfold gap that leaves SpaceX’s AI unit little room to price its way out of the hole.

• $6.36 billion: the AI unit’s full-year 2025 operating loss on $3.2 billion of revenue.
• $60 billion: what SpaceX paid in stock in June to acquire Cursor, now folded into that same unit.
• $2.47 billion: the AI unit’s loss in just the first three months of 2026, triple its quarterly revenue.
• 20 times: roughly how much cheaper DeepSeek’s lowest token pricing runs compared with xAI’s cheapest Grok tier.

A Near Monopoly That Still Loses Money

Space, the rocket-launch business, sits opposite AI in one sense and matches it in another. SpaceX has flown roughly 650 missions, reusing boosters on 85% of them, and carried an 80% share of everything launched into orbit worldwide in 2025, according to the registration statement SpaceX filed with the Securities and Exchange Commission (SEC).

There is not much room left to grow from there. Space spent $1.28 billion to generate $619 million in first-quarter revenue and still lost $662 million, a gap that would persist even if the segment doubled its output and turned every additional dollar into pure profit, an unlikely outcome for a business already running near the technical ceiling of the global launch market.

Execution risk has not gone away either. Two launch aborts within four days earlier this year briefly rattled the stock, a reminder that even an 85% reuse rate does not make the rocket side of the business risk-free.

A Pattern Wall Street Has Seen Before

The gap between what a company’s backers insist it is worth and what its financial statements show is familiar territory for institutional money. Theranos, the blood-testing startup, drew billions of dollars from sophisticated investors before its core technology was shown not to work. Wirecard, the German payments company, kept the confidence of major shareholders for years until a Financial Times investigation uncovered a hole in its accounts that pushed the firm into insolvency in 2020.

SpaceX flies real rockets and serves real broadband customers, a different footing than either scandal. What connects them is the pattern Morgan Stanley is now betting against: a market that stops taking a growth story on faith eventually forces a reckoning, and that reckoning has caught confident backers off guard before.

The stock’s own short life shows how fast sentiment can swing. The prospectus became publicly available in May, an amended version followed in early June, and shares opened trading days later at $160.95. They climbed to $201.80 within two weeks, then gave back more than $86 in barely six weeks, a round trip few trillion-dollar companies complete that quickly.

What Happens When Insiders Can Sell

Jonas’s note points to the mechanism that could push shares toward $100: a lockup expiration next month that frees early investors and employees to sell for the first time since the IPO. SpaceX stock already rallied 3.7% as a 911 million share unlock neared, an early sign of how sensitive the stock is to new supply hitting the market.

If enough of those newly tradable shares get sold at once, the stock could test $100 well before anyone resolves whether Grok, X and Cursor are worth zero, billions, or somewhere in between. Morgan Stanley is telling clients to buy that dip. Whether the call ages well or joins the pile of confident notes once written about Theranos and Wirecard depends on numbers nobody, including Jonas, can see yet.

For now, the distance between SpaceX’s rocket business and a market verdict that its AI arm is worthless comes down to $15.07 a share.

Frequently Asked Questions

What is Morgan Stanley’s price target for SpaceX stock, and what does it imply?

Morgan Stanley analyst Adam Jonas has set a $300 price target with an Overweight rating, a figure that would put SpaceX’s total equity value near $4 trillion, roughly double where the stock closed on July 24.

What businesses make up SpaceX’s AI segment?

The AI division houses Grok, the xAI chatbot; X, the social platform formerly called Twitter; and Cursor, the coding tool SpaceX acquired for $60 billion in stock in June 2026. Combined, those platforms reported 350 million daily posts and 550 million monthly active users last year.

Has SpaceX stock actually fallen to $100 a share?

No. Shares closed at $115.07 on July 24, still $15.07 above the $100 level Morgan Stanley calculates would imply zero value for the AI unit, though the stock has already lost more than a quarter of its value since the IPO.

Why does SpaceX’s share lockup expiration matter?

SpaceX barred pre-IPO investors and employees from selling shares for a fixed window after the offering. That window closes in the coming weeks, and Jonas specifically flags the added supply as the likely reason shares could test $100 before the AI unit’s value is settled either way.

Is Starlink profitable on its own?

Yes. The connectivity segment, dominated by Starlink, posted a $4.42 billion operating profit on $11.39 billion of revenue in 2025, an operating margin near 39%, making it the only one of SpaceX’s three reporting segments consistently in the black.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. Shares of newly public companies like SpaceX carry elevated volatility and risk; consult a licensed financial advisor before making investment decisions, and figures are accurate as of July 27, 2026.