Elon Musk Scales Back Solar Power Efforts on Earth

Has Elon Musk turned away from Tesla’s Master Plans, the electrified economy, and solar energy as we know it? The latest SpaceX IPO filing hints that he might have.

For those unfamiliar with the Musk-verse: Tesla has introduced four Master Plans over the years, each varying in specifics but consistently focusing on the electrification of the economy. Musk emphasized this in his first edition: “the overarching purpose of Tesla motors…is to help expedite the move from a mine-and-burn hydrocarbon economy towards a solar electric economy.”

However, Musk’s company xAI has recently shifted towards traditional energy sources, utilizing numerous unregulated natural gas turbines to power its data centers and planning an investment of an additional $2.8 billion in fossil fuels, reinforcing their role in AI operations.

This change is notable for a businessman who has built his empire on clean energy — and who candidly encourages collaboration among his companies. SpaceX has invested $131 million in 1,279 Cybertrucks, while xAI has committed $697 million over the last two years to Tesla Megapacks, the company’s large-scale battery storage solutions for peak management. Still, xAI has not yet acquired a significant number of solar panels from Tesla.

Solar energy isn’t absent from the SpaceX filing; rather, it is strictly centered on space, which the company promotes as the future for energizing data centers. While terrestrial solar receives a few mentions — not as a power source for xAI’s data centers, but to illustrate how much SpaceX believes space-based solar will excel.

It’s widely recognized that Musk and other Silicon Valley leaders are fascinated by solar power harnessed in space. SpaceX claims that solar arrays in space can generate “more than five times the energy” of ground-based ones due to uninterrupted sunlight. As AI data centers face challenges on Earth, executives like Musk are considering large server arrays in space, buoyed by that constant sunlight. Hammer, meet nail.

Even if SpaceX successfully lowers costs for launching data centers into orbit, the financial aspect remains intimidating. The power expenses for Starlink satellites are significantly higher than those of typical terrestrial data centers, and shielding components from the harsh environment of space will be neither simple nor cheap. It’s also unclear whether AI training can be evenly distributed across multiple satellites, leaving many AI operations confined to Earth. SpaceX confronts numerous challenges ahead.

Musk likely views xAI’s current data centers as temporary measures, believing that once SpaceX achieves the ability to launch gigawatts of servers into orbit — likely just a few years away in his perspective — he’ll phase out ground-based operations, including the natural gas turbines, sidestepping any NIMBY concerns. The inherent risk, of course, is the potential for mistakes.

Musk’s concerns go beyond NIMBY issues. He seems to be apprehensive that the computational demands of AI will quickly outstrip our Earthly capabilities. Scattered throughout the SEC filing are references to “terawatt-scale annual AI compute growth,” which will require matching power. This is a staggering estimate when considering that all of the world’s data centers currently consume around 40 gigawatts.

This illustrates Musk’s “first principles” approach. At some point, he determined that the world would need an additional terawatt of computing power each year and reverse-engineered his assumptions from there. “We believe that third-party estimates on data center demand are constrained by practical supply limitations that exist in a terrestrial context and that the power shortage may be far greater than what research indicates,” the company asserts.

Is it feasible? Definitely, one might argue. However, it’s important to note that humanity currently consumes roughly 35,000 terawatt-hours of energy each year, which amounts to about 4 terawatts continuously. Energy demands have surged recently, and while AI is likely experiencing rapid growth, it could either continue to rise or plateau. We can’t predict this now, but Musk excels at identifying trends at inflection points and extrapolating from there.

This is where Musk’s challenges return to Earth. I may not be a rocket scientist, but transporting solar panels via a flatbed truck likely consumes less energy than launching them into space. Furthermore, solar panels designed for space will need to be manufactured on a massive scale. These challenges are not insurmountable but may serve as a distraction. We have barely begun to explore solar’s potential here on Earth.

The ideal shouldn’t hinder the achievable. There’s ample opportunity to improve conditions here on Earth while simultaneously pursuing our aspirations among the stars.

Just three years ago, Musk and his colleagues at Tesla released “Master Plan Part 3,” which meticulously outlined a “plan to eliminate fossil fuels.” A sensible starting point might be within xAI’s data centers.