Morgan Stanley: Robotics to Become Starlink's Largest Revenue Source
nashnova research
Morgan Stanley projects robots will generate 36% of Starlink's revenue by 2040, with autonomous-vehicle connectivity alone reaching $31.2 billion — a sign that Starlink is shifting from selling broadband to selling robot infrastructure.
Why could robots drive the bulk of Starlink's revenue?
The report forecasts robot customers will account for 11% of Starlink revenue in 2030, rising to 29% in 2035 and 36% in 2040.
Autonomous vehicles lead the charge: Starlink revenue from self-driving cars is projected to grow from $6.4 billion in 2035 to $31.2 billion in 2040.
This means → Starlink's future growth engine is not phone users or home broadband — it is millions of machines that need always-on connectivity.
What does the Cybercab antenna tell us?
Tesla's Cybercab has a Starlink phased-array antenna — hardware that auto-tracks satellite signals — built into its roof. Tesla has hinted at launching a ride-hail service in Dallas soon.
The roots go back to 2018: analyst Adam Jonas asked on an earnings call whether SpaceX could provide a secure data link for Tesla vehicles. Musk said he would consider connecting Tesla cars to the low-earth-orbit constellation.
In plain terms = Tesla needs Starlink for uninterrupted, everywhere connectivity for its autonomous fleet; SpaceX needs that fleet as a massive paying customer base. The hardware is now physically installed.
Why do robots need so much bandwidth?
Future robots must transmit camera and radar perception data, run on-board AI inference — letting the machine decide its next move locally — and coordinate in real time with other robots.
The report notes one Tesla stores energy equal to 3,000–4,000 iPhone 18 batteries. Robots have ample spare power to boost communication signals; the extra draw is modest relative to locomotion and compute.
This reflects a demand profile that goes far beyond "basic connectivity" — robots need high bandwidth, low latency, and 24/7 coverage, which is precisely Starlink's sweet spot.
What does 10 million robots in one city look like?
The report models a metro area with 10 million robots, each carrying 1 kilowatt of compute. Total edge-inference capacity — processing data on the device, not in the cloud — would reach 10 gigawatts.
Low-altitude drones and high-altitude robots are also counted as potential customers, possibly relaying data through mesh networks where machines route signals to one another.
This means → the robot network itself becomes a massive communications-infrastructure market, and Starlink is positioned to serve as its core connectivity layer.
How does Morgan Stanley value SpaceX?
Morgan Stanley maintains an Overweight rating on SpaceX with a $300 target price. Since early July the bank has raised its 2027 and 2028 EPS estimates by 80% and 60%, respectively.
The valuation uses a sum-of-the-parts method: Enterprise AI contributes $170 per share, Connectivity (including Starlink) $111, X & Grok $11, and Space $8. The model applies an 11.3% WACC and extends forecasts to 2040.
Put simply = more than half of SpaceX's value, in Morgan Stanley's view, comes from AI and connectivity, not traditional rocket launches. Whether robot deployment scales as projected — and how deeply Starlink actually penetrates robot communications — are the key variables that will make or break this valuation.
市场有风险,内容仅供研究参考,不构成投资建议。
