Morgan Stanley’s SpaceX Math: Enterprise AI Is 55% of a $300 Target
TREE NEWS reports: Morgan Stanley reiterated an Overweight rating on SpaceX with a $300 price target, implying 83.6% upside from a $163.42 share price. The headline number matters less than its composition: in the bank’s sum-of-the-parts model, enterprise AI alone accounts for $165 of the $300 target — roughly 55% of total value — against $118 for connectivity, $8 for space, and $8 for X and Grok. Valuation is dated June 30, 2027, with forecasts extending to 2040, a weighted average cost of capital of 11.2% and a cost of equity of 11.9%. The enterprise AI piece is discounted by half to reflect execution risk.
The Market Is Pricing Rockets, Not Inference
The core argument is that investors recognize the observable business — rockets, launch pads, ground stations, data centers, satellites, terminals, spacecraft, solar panels, turbine blades, semiconductors, spectrum and AI models — but not the inferred one. Morgan Stanley frames SpaceX as the most vertically integrated manufacturer and infrastructure player Wall Street has seen in three decades, spanning consumer, industrial and national security demand. Its unifying equation, i=erA², describes converting energy into intelligence through physical agents: a small amount of energy connected to a universal cluster of agents yields enormous intelligence. As agent density rises, so does the collective capability of the swarm — the bank’s analogy is a carpenter ant colony, where 15,000 individuals operate as a single intelligent cluster.
Distributed Inference Cloud: The Unpriced Asset
The report identifies the distributed inference cloud as the largest and least-priced growth vector. Sensitivity work suggests each additional gigawatt of nameplate compute capacity — valued at $50 per watt, at a 70% incremental margin, capitalized at 10x EBITDA — adds about $27 per share, or 15%–20% of the current price. Morgan Stanley models 4.9 GW of AI compute in fiscal 2027 against a company goal approaching 10 GW. On 2028 enterprise value/EBIT/growth multiples, SpaceX screens at 0.59x versus 4.60x for Tesla, 1.21x for Microsoft and 1.61x for Palantir.
What Would Re-Rate the Story
- Upside: faster Starship reusability, stronger Starlink capacity and direct-to-cell adoption, more new cloud orders, accelerating Cursor ARR, and cheaper or faster power delivery for AI infrastructure.
- Downside: slower reuse cadence, weaker Starlink subscriber growth, soft enterprise AI monetization, rising cost per watt, longer power timelines, greater financing needs and dilution, and regulatory delays.
For public-market investors, the practical read-through is that SpaceX’s AI narrative is being valued like an option rather than a business. If the distributed inference cloud begins to show contracted revenue and power procurement milestones, the re-rating could come first gradually, then suddenly — the pattern the bank expects in investor discussion over the coming year.
