Forget rockets: Morgan Stanley sets $300 SpaceX target based on a secret AI weapon

Morgan Stanley has initiated coverage of Elon Musk’s newly listed SpaceX with an Overweight rating and a $300 price target, arguing that the company’s secret AI weapon, a vertically integrated terrestrial‑plus‑orbital compute stack, could make it one of the most powerful infrastructure platforms of the AI era.

In a note, the brokerage firm says SpaceX now combines “near‑monopoly launch economics, the world’s largest LEO satellite network, and a fast‑scaling AI infrastructure business” into a single, integrated stack of real estate in orbit, global connectivity and compute capacity. The bank’s base‑case model projects SpaceX revenue rising from about $45 billion in 2026 to $319 billion in 2030 and an eye‑catching $3.3 trillion by 2040, with the largest upside tied to Starship, Starlink capacity, terrestrial compute and orbital compute.

Four KPIs will drive the stock over the next several years, the analysts argue: revenue per watt, cost per watt, cost per kilogram to orbit and Starlink subscribers or connected “nodes.” These metrics map directly to the four big debates in their coverage — whether Starship can deliver a step‑change in launch economics, whether Starlink can achieve broad adoption, whether SpaceXAI can lead on compute cost and time‑to‑power, and how far the company can push enterprise AI monetisation.


Also Read | SpaceX set to join Nasdaq-100 today. Will index entry fuel its next rally?

The secret AI weapon: SpaceXAI and orbital compute

Morgan Stanley’s thesis hinges less on rockets and more on what those rockets enable: a hybrid AI infrastructure that starts on Earth and then moves “off Earth.” The report highlights SpaceXAI’s COLOSSUS and COLOSSUS II clusters, which collectively provide around 1.0 GW of compute power and related data‑centre capacity, with the first COLOSSUS cluster reportedly brought online in 122 days and COLOSSUS II in just 91 days. The analysts model SpaceXAI adding 1.2 GW of compute capacity in 2026, 2.2 GW in 2027 and 3.7 GW in 2028, making “time‑to‑power” a direct driver of revenue growth and share gains.

On costs, they believe SpaceX can achieve “industry‑leading compute cost and time‑to‑power, first through terrestrial compute and vertical integration, and longer term through orbital compute as a meaningful driver of lower cost per watt.” A large portion of capex is directed towards Terafab, Solarfab and other vertical integration efforts, including blade/vane foundry and terrestrial communications infrastructure, to push infrastructure capex to roughly $4–$5 per watt (excluding chips), versus industry averages closer to $9 per watt.

The “secret weapon” is orbital compute. Morgan Stanley has built a bottom‑up model across launch, satellite hardware and compute payloads and estimates three generations of AI satellites, progressing from roughly 150 kW of total power and 2.1 tonnes of mass to about 913 kW of power and 6.1 tonnes per satellite. They model orbital compute deployments beginning in 2028 at 160 MW, reaching 2.7 GW in 2030, 21 GW in 2032, 111 GW in 2035 and 364 GW by 2040, with orbital becoming the majority of total compute capacity by 2032.

Crucially, Morgan Stanley argues that orbital compute could reach cost parity with current industry terrestrial compute by 2031 on an all‑in annualised basis. Their model implies orbital compute capex per watt falling from about $166/W in 2028 and $60/W in 2030 to $32/W in 2031, $15/W in 2035 and $9/W by 2040, driven by lower launch costs, cheaper satellite hardware and improved payload economics. Using a five‑year useful life, that translates to roughly $6.5 per watt per year — slightly below their internet team’s estimate of about $6.8 per watt per year for Blackwell‑class terrestrial data centres.

“In our opinion, scalability/time‑to‑power is the bigger constraint vs. pure cost,” the analysts write, noting that recent neocloud deals have supported premium pricing due to the scarcity of large, high‑end GPU clusters that can be accessed quickly. They argue that even if orbital compute is not immediately at cost parity, SpaceX may still be able to charge a “material premium” if it becomes “the most scalable AI infrastructure player available,” drawing a parallel with the company’s launch business, where prices have risen despite steadily falling internal costs.

Also Read | Not just Elon Musk, retail investors are losing money in Tesla and SpaceX crash. What should you do?

From rockets to revenue: Starship and Starlink as AI enablers

While the note’s headline is all about AI, Morgan Stanley is explicit that Starship remains the “main unlock” for SpaceX’s long‑term economics across space, connectivity and AI. The team assumes Starship becomes operational in the fourth quarter of 2026 and models launch costs per kilogram falling to roughly $500 by 2030, under $200/kg by 2035 and below $150/kg by 2040, versus historical averages of about $18,500/kg pre‑Falcon 9.

Starship’s reusability profile is central to that curve. So far, SpaceX has caught the Super Heavy booster three times and reused it twice, and the next milestone is physical recovery of the Starship upper stage, or “Ship,” which has already performed controlled ocean landings. Morgan Stanley expects Ship recovery “sometime before year‑end, potentially as soon as Flight 14,” but cautions investors to focus “less on the literal landing itself and more on the actual turnaround time required for Ship reuse,” as cadence will be the key indicator of cost progress.

By 2040, the analysts model roughly 6,000 Starship launches per year, implying about 16.5 launches per day across five pad complexes and 10 towers, with each Ship reused around 40 times and each booster about 130 times, supported by a fleet of roughly 170 Ships and 56 boosters. They stress that 75% to more than 90% of Starship launches are likely to be used internally between 2027 and 2040, meaning launch economics flow directly into Starlink V3 deployment, Mobile Gen 2 satellites, orbital compute and other infrastructure programmes.

On the connectivity side, Morgan Stanley frames Starlink as the “connectivity layer for every data‑transmitting device that needs reliable coverage beyond terrestrial networks.” With Starship‑enabled V3 broadband and Mobile Gen 2 satellites, they estimate saleable capacity rising about 2.5x in 2027 on just 35 Starship launches, nearly 19x by 2030 and roughly 600x by 2040, with V3 satellites offering more than 10x the downlink capacity per satellite and over 20x the capacity per launch versus current V2 Mini on Falcon 9.

Connectivity revenue is modelled to climb from $11.4 billion in 2025 to $120.6 billion in 2030 and $687.7 billion in 2040, driven by international consumer broadband, Starshield, enterprise, mobile and a wave of Starlink‑connected robots and other embodied AI devices. Morgan Stanley estimates a global robot base of about 2.2 billion by 2040, with Starlink penetration rates of 40% for autonomous vehicles, 33% for drones and eVTOLs and 20% for other robots, contributing roughly 34% of 2040 Connectivity revenue.

To reach broad adoption, the bank’s model assumes global household broadband penetration peaking near 11.5% in 2040 and mobile penetration at around 6%, while consumer and enterprise ARPU declines from roughly $92 per month in 2025 to $47 in 2030 and $36 in 2040 as international markets grow in the mix.

Enterprise AI: neocloud today, agents tomorrow

Morgan Stanley calls Enterprise AI “the largest opportunity” in its SpaceX model and breaks it into three layers: renting compute through neocloud agreements, selling managed AI infrastructure and tools, and delivering end‑to‑end enterprise applications. They forecast AI revenue rising from about $22 billion in 2026 to $190 billion in 2030 and $2.6 trillion in 2040, with enterprise use‑cases accounting for the majority of the segment over time.

Near term, neocloud dominates. SpaceX’s S‑1 filing disclosed a major Anthropic deal, and the company has signed an agreement with Google and reportedly entered talks with Reflection, which Morgan Stanley collectively estimates at roughly $28 billion of annualised revenue through 2029. “We expect neocloud to drive most revenue given tight compute supply,” the analysts write, adding that the company “expects to enter into additional services contracts” in reference to the Anthropic deal.

Longer term, the mix is expected to shift towards higher‑value software and workflow automation. Products such as Cursor, Grok Enterprise, Macrohard and agentic AI tools could push SpaceX deeper into coding, digital agents, robotics, autonomous operations and other AI‑enabled workflows. The key debate, Morgan Stanley says, is whether SpaceX can “combine compute, models, data, and distribution to capture more value per watt than neocloud peers.”

Their base case assumes 6.1 GW of average compute capacity in 2028 and about $101 billion of AI revenue, equal to roughly $16.6 per watt — down 9% from a 2027 estimate of around $18 per watt as supply normalises. They outline multiple paths to $100 billion of AI revenue in 2028, ranging from 4 GW at $25 per watt if supply remains tight to 8 GW at $12.5 per watt if broader industry capacity improves. Over time, they expect revenue per watt to “decline meaningfully” as the market moves towards equilibrium, but argue that software and services could help support higher value extraction than pure compute rental.

Valuation and risks

Morgan Stanley’s $300 price target is built on a 15‑year divisional DCF sum‑of‑the‑parts, with an 11.1% WACC and a 50% discount applied to the AI business, and triangulated against “high growth mega‑cap tech comparables.” At $300 per share, SpaceX would trade at about 25x 2028 sales, 56x 2028 EBIT and 36x 2028 EBITDA, while a growth‑adjusted EV/EBIT framework screens below the 25th percentile of a comp set of large AI enablers.

The bull case of $600 per share assumes faster execution across Starship, orbital compute and Terafab with AI representing more than 60% of valuation, while the bear case of $75 per share assumes Starship slips to 2029 and AI monetisation and deployment speed disappoint, leaving Space and Connectivity at roughly 90% of valuation. “Space is hard and future Starship and in‑space anomalies should be expected,” the analysts caution, flagging technology readiness, launch anomalies, regulatory and geopolitical exposure and governance risks linked to Elon Musk’s central roles and expected voting control after the IPO.

(Disclaimer: Recommendations, suggestions, views and opinions given by the experts are their own. These do not represent the views of The Economic Times)