Something Big Is Happening
Elon is back and Tesla and SpaceX are preparing to spend like it’s 2005 again
When founders leave the arena and return, they tend to do so with either scars or fire in their belly. In Elon Musk’s case, it appears to be both. Since returning from his stint in government there is a noticeable change in Elon’s messaging. The ambition feels massive again.
Tesla and SpaceX were founded in the early 2000s with narrow theses:
Tesla was about accelerating the transition to sustainable energy through electric vehicles and batteries.
SpaceX was about reducing the cost of access to space through a low cost launch platform.
Neither master plan contained footnotes about artificial general intelligence, global autonomous fleets (of cars or humanoid robots) or, especially, orbital data centers.
They didn’t need to.
The most underappreciated strategy in technology is not prediction, it’s capability building. If you assemble rare, compounding capabilities (e.g., manufacturing at scale, vertically integrated supply chains, electric powertrains, self-driving software, reusable rockets, satellite constellations, etc.) you create option value. You don’t need to foresee every opportunity. You simply need to be positioned when they appear.
Consider Tesla first. Tesla started with a single, focused bet: build a great electric car. To do that well, they had to master things that had never been done at scale. They designed high-density battery packs, electric powertrains, over-the-air software updates, and eventually, a vast network of sensors and cameras generating real-world driving data. None of that was built with self-driving, robotics or AI infrastructure in mind. But then it turned out the data from millions of vehicles was an asset that could be used to develop full self-driving. The capabilities required to solve autonomous driving are, it later turned out, the same capabilities required to build a distributed robotics network.
Now consider SpaceX. SpaceX started with an equally narrow thesis: make rockets reusable and dramatically cut the cost of getting to orbit. Achieving that required building vertically integrated rocket manufacturing, advanced propulsion systems, and eventually the logistics infrastructure to launch frequently and reliably. Starlink emerged not from a grand vision of owning global internet infrastructure, but from the natural extension of having cheap, frequent access to orbit. Now that capability stack looks increasingly strategic in ways no one anticipated in 2002. If AI workloads increasingly demand distributed, resilient and power-abundant compute environments, orbit begins to look less exotic and more essential. Nevertheless, the original thesis was just about rockets.
Here is the throughline: both companies spent twenty years building hard, physical and difficult-to-replicate capabilities. The world has now shifted in a direction that amplifies those capabilities rather than obsoleting them.
But now that we’re here, I’m predicting that we’ll see substantial increases in capital expenditure from both Tesla and SpaceX in 2026. Not marginal tweaks, but step-function expansions. These are once-in-a-generation platform shifts: autonomy at scale, AI-powered robotics and the industrialization of low Earth orbit. Musk has historically erred on the side of overbuilding into opportunity, so he is unlikely to underinvest.
Critics will warn of technology risk, execution risk and capital discipline. They will not be wrong to do so. But history suggests that when foundational technologies converge (energy, compute, labor saving inventions and transportation) the winners are those willing to place large bets.
The early 2000s versions of Tesla and SpaceX could not have predicted this moment. Fortunately, they built the kind of companies that don’t need to predict the future, only to recognize it when it arrives and spend accordingly.
Please “Like ♥️” this piece!
