SpaceX and xAI: Elon Musk Is Building an Engineering Behemoth in Pursuit of AGI

Elon Musk has never liked playing by the rules of classic corporate governance, and his new plan to merge SpaceX and xAI is the best proof. While skeptics talk about conflicts of interest, Musk sees the merger as the only way to survive the race for AGI. The idea is simple and unsettling at once: combine the most successful private space company in the world with a fast-growing AI startup. This is not just shuffling money from one pocket to another, but an attempt to build a closed loop for producing intelligence. SpaceX has what xAI badly lacks: enormous engineering capacity, experience with extremely complex physical systems and, importantly, a steady cash flow from government contracts and Starlink.

Context matters a great deal here. Recall that xAI started as a modest answer to OpenAI and is now building the world's largest supercomputer, Colossus. Running it takes more than Nvidia chips; it also needs specialized infrastructure, which SpaceX can build faster and more efficiently than any construction company.

On top of that, data from Starlink satellites and rocket telemetry is a unique pool of information for training models, one that neither Google nor Microsoft will have access to. Musk is effectively building a digital nervous system, with SpaceX as the body and xAI as the brain. It is a logical extension of the strategy we have already seen at Tesla with its FSD, only now the stakes have grown to planetary scale.

Against the backdrop of these reshuffles, Larry Ellison's Oracle is not standing idle either. The company is going to the bond market intending to raise $45 to $50 billion this year. Why does a software giant from the 90s need that kind of money?

The answer is one thing: clouds for AI. Oracle stopped being just a "database company" long ago and has become Musk's main ally when it comes to renting computing capacity. These billions will go toward building new data centers that will form the foundation for training next-generation neural networks.

Ellison and Musk both understand that in a world where each new model needs ten times more resources than the last, the winner will be whoever has more "concrete and iron."

Why does this matter right now? We are passing an inflection point at which pure software stops being an advantage. OpenAI can write brilliant code, but without access to cheap energy and endless server racks, it remains hostage to Microsoft. Musk is trying to break free of that dependence. If the SpaceX and xAI merger goes through, he will gain access to capital that lets him ignore venture investors. That sets a dangerous but remarkably effective precedent: an AI company backed by real assets in space and on the ground.

The link between Oracle and Musk's empire is getting tighter. While Oracle builds "digital factories," Musk is preparing workloads for them that once seemed like science fiction. This is not just business news; it is the formation of a new technological oligarchy, in which computing power becomes hard currency and the ability to build physical infrastructure is the main competitive advantage. We are entering an era in which the line between a software startup and heavy industry is finally disappearing.

Bottom line: Musk is making a final bet on vertical integration, turning xAI into part of his industrial empire. Can anyone compete with an AI that has its own rockets and satellite network?