Morgan Stanley gives SpaceX (SPCX.US) a new label: a “safety cushion” amid AI turmoil, with a target price of $300

According to Zhitong Finance APP, since last Friday, a debate over whether the development of cutting-edge AI models should be “proactively slowed down” has rapidly spread from the tech community to the capital market. On September 12, Anthropic’s CEO, Dario Amodei, released a three-step plan calling to “set the pace for frontier AI development.” OpenAI’s Sam Altman subsequently stated that he would adopt Amodei’s proposal for “independent evaluators with employee-level access,” and even Elon Musk, known for his aggressive stance, has unusually joined this camp. On the first trading day after the news fermented, global tech stocks took a sharp dive.

However, amidst the panic selling of AI concept stocks, Morgan Stanley analyst Adam Jonas delivered a rather unconventional judgment to clients: SpaceX (SPCX.US) may actually be one of the stronger defensive AI-related assets during this turmoil. In a research note published on September 15, Jonas reiterated his “Overweight” rating on SpaceX and a price target of $300, providing the following reasons.

Diversified Foundation: More Than Just an “AI Concept Stock”

Jonas made an honest statement at the beginning of his report. He wrote: “You can’t have it both ways.” The implication is that the vast majority of SpaceX’s capital expenditures and long-term profit growth are indeed driven by enterprise AI business; any substantial slowdown in AI trading is a headwind for this stock. But on the other hand, SpaceX’s revenue structure is not purely a bet on the narrative of AI computing power.

According to the company’s Q2 2026 financial report released on August 4, the number of Starlink global subscribers has reached 12 million, doubling from 6 million in the same period last year, with a net addition of 1.7 million in a single quarter. The continued expansion of enterprise and government clients has made Starlink SpaceX’s largest source of revenue. Jonas believes that the growth driver of this cash cow business comes from the market penetration of satellite communications itself, not from industry-wide AI capital expenditure cycles.

Meanwhile, SpaceX’s Starship project is also advancing at its own pace. According to the latest notice from the US Federal Aviation Administration, the 14th Starship test flight has been rescheduled to as early as September 22, launching from Starbase in Texas. This mission will attempt for the first time to enter orbit and deploy real Starlink satellites. More noteworthy is the 15th flight—if the 14th mission is completed successfully, SpaceX plans to conduct the next launch in less than 30 days, and during the 15th mission, will attempt to simultaneously catch the Super Heavy booster and Starship with the launch tower’s “chopstick” robotic arms. This will be a key validation step for a fully reusable system.

Jonas believes that the Starship launch schedule is a typical engineering-driven milestone, and is almost unrelated to the progress of AI model training or regulatory maneuvering.

Long-term Option: Orbital AI and Robotics Collaboration

If Starlink and Starship constitute SpaceX’s “non-AI fundamentals,” then the space computing project is, in Jonas’s eyes, a long-term option. In early 2026, SpaceX applied to the FCC for permission to launch up to one million data center satellites. In July, the company officially disclosed its orbital data center project called “Starmind.” The first AI satellite, AI1, has a peak computing power of about 150 kW, equipped with a 70-meter solar sail span and a liquid-cooled radiator system. SpaceX is also working with Nvidia (NVDA.US) to customize a space-optimized version of the Vera Rubin NVL72 architecture-based AI computing system for orbital environments, with the first batch targeted for launch in Q4 2027.

Jonas admits that this vision probably won’t be realized at scale until after 2030; in essence, it is “a very long-duration option.” However, he also notes that the current market has assigned almost no valuation to the collaborative synergy between SpaceX and Tesla (TSLA.US) in robotics and domestic chip manufacturing.

According to public information, the two companies are jointly advancing the “Terafab” semiconductor manufacturing project, aiming to integrate logic chips, storage, packaging, and testing into one facility, supplying chips for Tesla’s Optimus humanoid robot and Cybercab, while also providing components for SpaceX’s developing space data centers. Tesla has acquired a stake in SpaceX, and the two sides have signed agreements covering joint projects such as Terafab and “digital Optimus.” Musk himself has described this relationship as “mutually dependent” and warned that “without Terafab, Tesla’s ability to ramp up Optimus production will be constrained.”

Financial Flexibility: Cash Rich, Borrows As Needed

At the balance sheet level, Jonas particularly emphasized SpaceX management’s initiative in capital allocation. By the end of Q2, the company held about $100 billion in total cash and around $6 billion in net cash.

Jonas expects the company will raise significant funds via the debt market in the future, pushing the debt/EBITDA ratio to 1.5–2 times, but crucially, both the pace and scale can be flexibly controlled by management—“It only spends what it can borrow, and only spends if the business plan works.” After the IPO, SpaceX has obtained investment-grade credit ratings from the three major international rating agencies and has issued $20 billion in short-term unsecured notes to repay bridging loans due next year. This financing capability creates an extra buffer in an AI slowdown scenario.

Valuation: Market Has Barely Priced in the AI Business

From a valuation perspective, Jonas estimates that about three-quarters of SpaceX’s $2 trillion valuation comes from its space and communications business. Excluding the AI business, the corresponding stock price is about $125—meaning the space business is valued at about $8 per share, satellite communications at $118 per share—and at current prices, the AI business (enterprise AI plus X and Grok) is only accorded about 10% of implied value, about $17 per share (based on its $183 AI valuation). As early as July, Jonas posited that “SpaceX at $100 means the AI business is valued at zero.”

He further performed a sensitivity analysis: for every 1GW increase in nameplate computing power (valued at $50 per watt, 70% incremental profit margin, and a 10x EBITDA capitalization), it can add about $27 to the per-share value, equivalent to 15%–20% of the current stock price. Morgan Stanley’s AI computing power forecast for fiscal 2027 is 4.9GW, while the company’s own goal is “close to 10GW.”

The “Musk Factor”

The fifth reason Jonas gives returns to Musk himself. In his report, he writes that in two decades of observing this CEO’s management style, “he is not the type to let a good crisis go to waste.” Tesla’s performance during the “production hell” of 2018–2019 and the COVID-19 pandemic of 2020–2021 is seen by Jonas as evidence of the company’s ability to achieve both operational and stock price breakthroughs under high-pressure environments.