Tech • AI • Robotics • Game
Why Did He Dodge the Tesla-SpaceX Merger Question?
9/10TL;DR
Morgan Stanley argues that the next major AI wave will be in the physical world, with Tesla and SpaceX positioned as complementary players in robotics, autonomy, connectivity and distributed compute.
KEY POINTS
Physical AI as an economic shock
Adam Jonas said the fusion of AI and robotics could multiply global output by 8x to 10x before today’s workers retire. With global GDP currently around $180 trillion in 2025, that would mark an unusually fast acceleration versus past economic expansions, reflecting a shift from human-limited labor to machine-augmented intelligence operating in the real world.
Tens of billions of robots
Jonas said daily life could include tens of billions of robots within the next 10 to 20 years. He described humanoids as only one layer of a much broader robotics stack that would also include low-altitude robots, terrestrial machines, autonomous vehicles, mobile systems and industrial automation.
Humanoids matter, but are not the whole market
In Morgan Stanley’s long-range modeling through 2050, humanoid robots still reach the billions in some scenarios. Jonas said the humanlike form factor attracts attention, capital and talent, but it represents only one category among thousands of machine types expected to carry embedded AI into homes, factories, logistics networks and transport systems.
Tesla and SpaceX seen as complementary
Jonas described Tesla as a manufacturing and data-collection layer that can build robots and scale intelligence into the physical world. He cast SpaceX as a connectivity and AI infrastructure layer with its own advanced manufacturing capability. Rather than directly address merger speculation, he said investors should expect continued cooperation because the relationship appears increasingly deterministic.
A shared mission around intelligence efficiency
The strategic overlap was framed as a drive to convert energy into intelligence at scale, maximizing intelligence per watt, dollar and second. In that view, robotics, autonomy, communications networks and computing infrastructure are converging into a single industrial race over how cheaply and quickly intelligence can be deployed.
Inference moves to the edge
Jonas argued that future AI cannot rely solely on centralized data centers. He said edge inference will become essential, comparing it to neurons in a brain, with robots and machines processing tasks locally while connected to broader networks. He described using a large data center GPU for routine inference as inefficient, likening it to using a Ferrari to pick up milk.
Machines without AI face faster obsolescence
Any machine that can carry an AI inference computer likely will, Jonas said, while equipment that cannot may face a sharply shorter useful life. The implication is that intelligence is becoming a standard industrial input, not a premium feature, and products across transport, manufacturing and consumer hardware may be redesigned around embedded compute.
Autonomous driving seen as effectively solved
Jonas said autonomous cars are no longer a theoretical challenge but a solved modality in practical terms, pointing to Waymo’s driverless service in Phoenix in 2023 as the key milestone. He acknowledged that safety performance still needs improvement, but argued that once insurers begin offering discounts for letting cars drive themselves, wider public adoption could accelerate and open the door for other robotic form factors.
Tesla’s current position in embodied AI
Jonas said Tesla, together with SpaceX, represents one of the strongest Western chances to keep pace with or eventually surpass geopolitical rivals in robotics and physical AI. He cited a hands-free trip in a Tesla from Westchester to Midtown Manhattan as an example of how far real-world autonomy has advanced, even as Chinese robotics companies remain formidable competitors in adjacent categories such as robot dogs.
CONCLUSION
The central bet is that physical AI will spread from cars into nearly every class of machine, turning robotics, connectivity and inference into core economic infrastructure. In that landscape, Tesla and SpaceX are increasingly viewed as linked pillars of a broader industrial AI system.
Ask a question