Morgan Stanley Prices SpaceX Stock as an AI Bet

Morgan Stanley analyst Adam Jonas kept a $300 target on SpaceX on Sunday, arguing the Nasdaq-listed rocket company is still being priced as if its AI unit were almost free. Shares of Space Exploration Technologies, ticker SPCX, closed Monday, October 5, at $171.09, up 7.63% from Friday’s $158.96, after his note titled “Cheap and Getting Cheaper” hit the tape.

The Overweight rating has not moved since July. What moved is the yardstick. Jonas is no longer asking investors to pay up for launch. He is asking them to stop filing a $2 trillion-class name under aerospace while the leftover stub is compute, Grok, and a Washington rebrand called Super Intelligence.

The Note Treats Rockets as Already Priced

At the $159 level in the Sunday note, SpaceX screens expensive on the headline multiple, about 30 times estimated 2028 EV/EBIT against roughly 16 times for a group of mega-cap AI enablers. Factor in year-over-year growth and the comparison flips. SpaceX trades at about 0.3 times 2028 EV/EBIT-to-growth, around 40% below the group’s 0.5 times median.

Even at $300, that growth-adjusted multiple would be only about 0.6 times, in line with Amazon at 0.5 and below Alphabet at 0.7 and Meta at 0.8. Jonas called the valuation an “and” problem rather than an “or” problem, because launch, Starlink, and compute have to be weighed together. He still told clients to build positions before Starship Flight 15.

THE MONDAY CLOSE

• The print: SPCX finished at $171.09, a $12.13 gain, about 75% below the $300 target.
• The range: The 52-week band runs from $104.83 to $225.64, and the stock has churned near $150 since early August.
• The debut: SpaceX priced at $135 on June 11, 2026, then closed its first session at $160.95, up 19.2%.
• The raise: The IPO first brought in $75 billion, then a record $85.7 billion Nasdaq listing after the overallotment.

First-day volume surpassed 500 million shares on an initial cross of 58 million, and the close pinned the company at a $2.1 trillion market value. A 20% retail slice, about $15 billion, was large enough to be its own IPO. SpaceX’s S-1 showed $18.67 billion of 2025 revenue. Jonas still models $319 billion by 2030 and $3.3 trillion by 2040. Morgan Stanley co-managed a SpaceX offering in the past year and has taken investment-banking fees from the company.

The $32 the Market Pays for AI

The leftover math is the whole argument. At $159, Jonas wrote, the stock already pays for most or all of SpaceX’s space and connectivity businesses, which Morgan Stanley values at $127 a share. That leaves about $32 for AI, or roughly 3 times 2028 sales, a multiple he compared with neocloud names.

The bank’s sum-of-the-parts model is harsher on that stub, and much richer. It puts $165 a share on enterprise AI after a 50% execution discount, $118 on connectivity, and $8 each on external launch and on X and Grok, then rounds $299 to $300. Those two views are not the same number. One is what the tape is paying. The other is what Jonas thinks the stack is worth if the AI slice is real.

MORGAN STANLEY’S $300 STACK

Compute pricing is the swing factor he named. Visible Alpha consensus, as cited in the note, assumes $17.60 per watt across 4.1 gigawatts. Each extra $10 per watt would add more than $40 billion of revenue with no change in deployed compute. Jonas put SpaceX’s recent short-term neocloud contracts at $30 to $50 per watt, which is the same cash-collection fight already enriching the chip and power owners of the boom. He wrote those variables can swing even next year’s revenue by a multiple.

Forty Clients and an Empty Room

The other tell in the note is demand, or the lack of it. Jonas said he asked a room of 40 clients last week who owned SPCX and got silence. He added that it has been “really, really quiet on SPCX lately,” and that lockup expiries “have largely proved to be a non-event.”

not a single hand went up

Adam Jonas, equity analyst, Morgan Stanley Sunday note

That empty room is why the $300 reprint moved the stock. A target that has sat still since July only bites when the shares have gone sideways and the holders are not in the room. Common worries he listed, Grok versus other frontier models and spectrum for Starlink Mobile, are in the price, in his view. A drop to $100 inside 12 months would, he said, take a clear slowdown in AI progress, a severe Starship setback, or a material dilutive event.

The skeptical read on the same facts is simpler. A bank that helped take the company public is talking up a name that has not held its first-day pop, while the 52-week high of $225.64 still sits well above Monday’s close. Jonas’s own client poll cuts against the idea of a crowded long. It does not settle whether the $165 AI line can earn its keep.

Musk Files SpaceX as Super Intelligence

The same weekend, Elon Musk did the category switch in public. On October 4 he wrote that “SpaceX is a super intelligence company,” then posted “No more AI. SI. It’s better.” Asked whether SpaceXAI would become SpaceXSI, he replied, “Yes, we will make that change.” SpaceX had bought xAI in February 2026 and folded the brand into SpaceXAI in July. The X account had not flipped when the rename was first described, and no timetable was given.

The prompt was political. President Donald Trump signed Executive Order 14434 on Super Intelligence on September 29, and the White House fact sheet told agencies to replace Artificial Intelligence with Super Intelligence in official correspondence, sites, and non-statutory documents. On October 4 he announced a Super Intelligence Force led by Director of National Intelligence Jay Clayton, with FTC Chair Andrew Ferguson, OPM Director Scott Kupor, and Emil Michael, undersecretary of war for research and engineering, and said it would report to him and Chief of Staff Susie Wiles.

FROM XAI TO A FEDERAL LABEL

1. February 2026: SpaceX acquires xAI and folds the model shop into the public company that later lists as SPCX.
2. June 12, 2026: SPCX opens at $150 and closes at $160.95 on Nasdaq and Nasdaq Texas.
3. July 2026: Jonas initiates at Overweight and $300; Musk’s AI unit is branded SpaceXAI.
4. September 29, 2026: Trump signs the order that puts Super Intelligence into executive-branch usage.
5. October 4, 2026: Musk claims the SI label, Trump names the Super Intelligence Force, and Jonas reprints $300.
6. October 5, 2026: SPCX rallies 7.63% to $171.09.

A rename does not change Grok’s scores or a booster’s catch. The joke that followed Musk’s posts treated SpaceXSI the way people still say Twitter. The tape still has to decide whether the federal vocabulary is a policy tailwind for a compute vendor or just a new coat of paint on a launch company that already had a $300 target in July.

When Does Starship Flight 15 Hit the Tape?

Jonas’s near-term calendar is now an AI calendar. Starship Flight 15 is slated for late October or early November, and he wrote that a ship catch “could be the biggest positive catalyst since the IPO.” Third-quarter results, also due in late October, are where he wants the first clean look at Cursor and Grok Bot economics. Flight 16 is penciled in before year-end, with Grok 4.8, 4.9, and 5.0 and more neocloud contracts on the same list.

THE CATALYSTS IN THE SUNDAY NOTE

• Flight 15: Late October or early November, with a ship catch framed as the largest positive since the June listing.
• Third-quarter results: Late October, aimed at Cursor and Grok Bot economics rather than launch counts alone.
• Flight 16: Another Starship attempt before year-end, stacked on top of the catch attempt.
• Grok drops: Versions 4.8, 4.9, and 5.0 listed as sentiment support for the AI stub.
• Neocloud contracts: More short-term compute deals at the $30 to $50 per watt prints he already cited.

If the leftover $32 is the mispricing, a catch no longer just proves reusability. It becomes the event that lets a growth-adjusted AI multiple stick. A scrub, a failed catch, or a Grok miss now hits the same ticker that still prints under Industrials in some market data. That is the second-order change in the Sunday note, and it is why a reprint of an old target could still add $12.13 in a day.

SpaceX Would Still Need Outside Capital

The July initiation that first set $300 already warned that the model is capital-hungry. Jonas saw a path to launch costs of about $500 per kilogram by 2030 and below $150 per kilogram by 2040 if both Starship stages reuse quickly. He also wrote that SpaceX could need $300 billion of annual capital spending by 2031, might not print positive free cash flow before 2035, and could require $84 billion a year of outside capital from 2027 through 2034. The October note did not retire that problem. It listed funding needs beside slower reuse, weaker enterprise AI sales, higher cost per watt, longer time-to-power, and regulatory delay.

So the $300 call and the $100 path live on the same page. The bull case is that $32 a share is too little for a neocloud that already rents at $30 to $50 per watt, and that Super Intelligence policy plus a Starship catch will close the gap to $165 of AI value. The other case is that a company still working through Starship tests, spectrum, and model quality will have to sell more paper to fund the stack that is supposed to justify the multiple.

Monday’s close at $171.09 only reopened that argument. Flight 15 is still on the calendar, the SpaceXSI name is still a promise, and Jonas’s $300 target is still the same number he wrote in July, waiting on compute to do the work rockets already did in the model.

Disclaimer: This article is news reporting and analysis of analyst research, public-company trading, and government announcements, and it is for information only. It is not investment advice, a recommendation to buy or sell SPCX or any other security, or a forecast of future returns. Readers should consult a licensed financial adviser or other qualified investment professional before acting on any price target, valuation multiple, or catalyst described here. Share prices, ratings, and corporate plans reflect the cited notes and market prints as of the dates given and can change without notice.