SpaceX is now being valued less like a rocket company and more like an AI infrastructure bet. Morgan Stanley's warning is simple: if the stock falls to $100, investors are giving the AI story no credit at all.
Morgan Stanley analyst Adam Jonas has put a hard number on the anxiety building around SpaceX. The Fly reported on July 24 that Jonas said many investors now expect SpaceX shares to hit $100 as the company's first post-IPO lockup restrictions begin to lift. At that price, according to Morgan Stanley's sum-of-the-parts work, the market is paying for launch, Starlink and Starship, while assigning nothing to AI.
That's a sharp turn for a company that priced its IPO at $135 on June 11 and began trading on the Nasdaq on June 12 under the ticker SPCX, according to SpaceX's own IPO release. The shares opened at $150, ran as high as $225.64 soon after listing, and were trading around $113 on July 24, based on StockAnalysis data. So the question isn't abstract anymore. You can see the repricing on the screen.
Jonas isn't backing away from the bigger call. Morgan Stanley initiated coverage earlier this month with an Overweight rating, a $300 base-case price target, a $600 bull case and a $75 bear case, as Saxo and other market outlets reported from the bank's note. That range is enormous. It should be. SpaceX is asking public investors to value rockets, broadband, AI compute and orbital data centers before several parts of that plan are fully proven.
The $300 case isn't really about Falcon 9 launch revenue. Morgan Stanley's analysts described SpaceX as one of the few platforms that can connect orbital real estate, global connectivity, and compute capacity - plus the data to tie all of it together - into one infrastructure stack. That's the bet. Launch gets payloads up. Starlink provides the network backbone. From there, xAI and Colossus supply the compute layer on the ground, with orbital compute pushed further out.
Frankly, you shouldn't treat that as a normal aerospace valuation. It isn't one.
The AI contracts are real, but so is the cash drain
The strongest part of the AI story is that it now has signed customers attached to it. Reuters reported in June that Google agreed to pay SpaceX $920 million a month from October 2026 through June 2029 for access to about 110,000 Nvidia GPUs, CPUs, memory and related components. TechCrunch reported the same terms from SpaceX's regulatory filing and noted that the Google deal followed a May agreement with Anthropic.
That Anthropic contract is larger. SpaceX's filing, as covered by TechCrunch and Morningstar, said Anthropic agreed to pay $1.25 billion a month through May 2029 for compute from the Colossus data center complex near Memphis, Tennessee. Together, those two deals point to roughly $26 billion in annualized revenue if they run as described. That's not a slide-deck fantasy. It's a commercial anchor.
But the costs are not small either. TechCrunch reported from SpaceX's IPO filing that xAI lost $6.4 billion from operations on $3.2 billion of revenue in 2025. In the first quarter of 2026, the AI segment's capital expenditures reached $7.7 billion. Morningstar reported that xAI posted an operating loss of $2.47 billion on $818 million of revenue in that same quarter.
That's the part investors can't wave away. SpaceX has a real compute business forming, but it also inherited the spending profile of a frontier AI lab. Morgan Stanley's earlier note, cited by Saxo, forecast no free-cash-flow-positive year before 2035 and average external capital needs of about $84 billion a year from 2027 through 2034. The company may be building something unusually valuable. It may also need an unusual amount of outside money to get there.
The lockup makes the valuation test immediate
The calendar now matters almost as much as the thesis. SpaceX said on July 20 that it will report second-quarter results after market close on August 4. Reuters previously reported, via Investing.com, that 911.5 million shares held by employees and some early investors can become eligible for sale on the second trading day after that first quarterly report. With the earnings date now set, that points investors to August 6.
Eligible doesn't mean sold. But when a stock has already fallen below its IPO price, the existence of that much possible supply changes the conversation. Reuters put the eligible block at about $123 billion when SpaceX was trading near $135 on July 16. At $113, it is still a very large number coming toward a market that has already turned cautious.
The next few days give shareholders two tests at once. SpaceX's delayed Starship flight attempt was pushed to July 24 because of weather, according to Investors Business Daily, after an earlier abort tied to Raptor engine issues. Then comes the August 4 earnings report. If investors hear more about Google, Anthropic and Starlink growth than they do about xAI cash burn, Jonas's $300 case gets more oxygen. If the spending overwhelms the contracts, $100 stops sounding like a scare line and starts looking like the next level to test.
SpaceX's problem is not that investors don't understand the story. They understand it well enough. At $300, you're paying for the idea that rockets, Starlink and the entire AI compute stack belong inside one company. At $100, you're saying the rockets and Starlink are enough, and the rest still has to prove it deserves a price.
Also read: Kimi K3 cracked chip stocks and split Trump's AI advisers in the same week • CATL's energy storage business grew 88% in six months as AI data centers become its next major market • The London Stock Exchange Is Rebuilding Itself for AI Agents That Never Stop Trading
