xAI Abandons Solar: Natural Gas and Orbital Data Centers

On May 23rd, TechCrunch published an analysis that captures what many in the industry already suspected: xAI, Elon Musk's artificial intelligence company, has decisively committed to natural gas as its primary energy source, while SpaceX explores data centers in orbit. The "solar-electric economy" that Musk promised less than a decade ago has been pushed to the background, at least as far as his own companies are concerned.

The fact is striking precisely because of who is behind it. For years, Musk used Tesla and SolarCity, later absorbed by Tesla, as central pillars of his thesis on energy transition. That his AI company now chooses gas and his aerospace company looks to space to solve the computational consumption problem says quite a bit about real priorities when scaling infrastructure rapidly.

Natural gas: the pragmatic choice nobody wanted to admit

xAI's data centers, necessary to train and serve models at scale, consume electricity in quantities that conventional power grids struggle to absorb quickly. Solar energy, given its intermittent nature and the deployment timelines for industrial-scale installations, doesn't align well with the urgency characterizing the current AI race.

Natural gas offers what solar cannot guarantee in the short term: dispatchable capacity, meaning electricity available when needed and in the quantity needed. It's a pragmatic decision, but it contradicts the public narrative Musk has maintained for years about the inevitability of renewable energy.

This isn't unique to xAI. Amazon, Microsoft, and Google have spent months reviewing their net-zero commitments under pressure to scale AI infrastructure. The difference is that these companies have communicated it with more rhetorical care. xAI apparently hasn't felt that need.

Orbital data centers: engineering or distraction?

The other focus of TechCrunch's analysis is SpaceX's obsession with orbital data centers. The idea has surface-level logic: in orbit, solar energy is constant and abundant, without day-night cycles or atmospheric losses. Solar panels in space capture four to eight times more energy per square meter than on Earth's surface.

But the challenges are considerable. Communication latency between an orbital data center and users on the ground is non-trivial. The cost of launching hardware to space, maintaining it, and eventually replacing it remains high even with Starship reducing the price per kilogram to orbit. And heat dissipation in a vacuum, one of the most complex problems in space engineering, becomes a serious obstacle when talking about GPU racks running at full capacity.

That said, SpaceX exploring this direction makes strategic sense: if launch costs keep falling and computational demand keeps growing, space's energy advantage could eventually offset operational difficulties. But "eventually" doesn't mean in the next two or three years.

Why this matters beyond Musk

The energy shift at xAI and SpaceX reflects a tension affecting the entire industry: industrial-scale generative AI has an energy appetite that clashes head-on with the sustainability goals these same tech companies had publicly set.

For those building on others' infrastructure, whether AWS, Google Cloud, or Anthropic's own API, providers' energy decisions indirectly affect the carbon footprint of their products. For those designing their own data centers or evaluating compute providers, xAI's example is a reminder that statements of intent and engineering decisions don't always align.

The relevant question isn't whether Musk is an energy hypocrite, a debate as old as Tesla itself, but whether the rest of the industry will be more transparent about the real trade-offs involved in scaling AI at this pace.

---

Editorial perspective: xAI's quiet abandonment of solar is, above all, a symptom that the AI industry hasn't yet solved its underlying energy problem. That the temporary solution is gas and the future solution is space reveals more about current limitations than about any coherent strategic vision.