SpaceX and Tesla’s AI linkup is getting deeper. Here’s what Morgan Stanley sees

Morgan Stanley has noted that Tesla and SpaceX are strengthening their technological and operational bonds, as both companies venture into the realm of practical AI. This alliance focuses on converting energy into intelligence. Both Elon Musk-led corporations share resources, talent, and infrastructure to achieve this shared goal. SpaceX President Gwynne Shotwell's comments before the company's IPO underline the synergies between the two entities.

In their S-1 filing, Morgan Stanley identified 71 separate references to Tesla within SpaceX's filing, 34% of which highlighted direct collaboration between the two companies. A further 41% of the references were connected to governance matters and related-party transactions. The analysts identified numerous areas where the two companies intersect, including a joint chip manufacturing facility known as Terafab, under construction in Texas alongside Intel, with an estimated cost surpassing $119 billion.

Additionally, SpaceX buys substantial quantities of Tesla batteries and vehicles, such as $506 million worth of Megapacks and $131 million worth of Cybertrucks in 2025, to support its Colossus compute cluster.

The two companies are also jointly developing an agentic platform called Macrohard. Each is also separately aiming to achieve 100 gigawatts of annual, vertically integrated solar manufacturing capacity within the U.S. SpaceX supplies Tesla with compute resources, projected to rise to 4.9 gigawatts by the end of 2027 and 15.3 gigawatts of terrestrial capacity by 2031.

Starlink satellite connectivity, which Tesla is integrating into its vehicle lineup starting with Cybercab, is another area of cooperation. SpaceX also provides Tesla with robots that function as data-collecting nodes and distributed compute infrastructure, alongside energy storage technology and manufacturing expertise. The analysts pointed out that Tesla has more than six times the workforce of SpaceX, with more than 135,000 employees.

As large language models (LLMs) near their performance ceilings, Morgan Stanley anticipates increased investor interest in underpenetrated areas like physical AI. While neither company significantly contributes to the pure LLM frontier model market, this scenario opens up a broad, unexplored potential in physical AI that the companies may tackle individually or, more likely, together.