Elon Musk Scales Back Solar Power Projects on Earth

Has Elon Musk distanced himself from Tesla’s Master Plans, the electrified economy, and solar energy? The latest SpaceX IPO filing may indicate he has.

For those unfamiliar with Musk’s story: Tesla has introduced four Master Plans over time, each differing in specifics yet consistently focusing on the electrification of the economy. Musk emphasized this in his first plan: “the main goal of Tesla Motors…is to aid in accelerating the transition from a mine-and-burn hydrocarbon economy to a solar electric economy.”

Recently, Musk’s xAI has shifted towards conventional energy sources, deploying various unregulated natural gas turbines for its data centers and planning a $2.8 billion investment in fossil fuels, underscoring their importance in AI operations.

This transition is notable for a businessman who has built his legacy on clean energy — and who actively promotes collaboration among his ventures. SpaceX has invested $131 million in 1,279 Cybertrucks, while xAI has spent $697 million in the past two years on Tesla Megapacks, Tesla’s large-scale energy storage units. However, xAI has yet to make a significant purchase of solar panels from Tesla.

Solar energy is not entirely absent from the SpaceX filing; it primarily focuses on outer space, where the company claims the future for powering data centers lies. While terrestrial solar does get a few mentions, it’s not positioned as a power source for xAI’s data centers, but rather to illustrate SpaceX’s belief in space-based solar outperforming ground-based alternatives.

It’s widely recognized that Musk and other Silicon Valley leaders are intrigued by solar power harnessed in space. SpaceX asserts that solar arrays in space can generate “more than five times the energy” of those on Earth due to continuous sunlight. Confronted with limitations on Earth, leaders like Musk are exploring large server arrays in space, bolstered by relentless sunlight. Hammer, meet nail.

Even if SpaceX successfully lowers costs for launching data centers into orbit, the financial challenges remain significant. Power costs for Starlink satellites are considerably higher than those of standard terrestrial data centers, and safeguarding equipment from harsh space conditions will be neither easy nor cheap. Furthermore, it’s uncertain whether AI training can be effectively distributed across multiple satellites, potentially confining many AI operations to Earth. There are numerous obstacles SpaceX must address.

Musk likely views xAI’s current data centers as temporary, anticipating that once SpaceX is able to send gigawatts of servers into orbit — which he presumably sees as a short-term goal — he will move away from ground-based operations, including the natural gas turbines, thereby sidestepping NIMBY challenges. However, the inherent risk lies in the potential for missteps.

Musk’s concerns extend beyond NIMBY issues. He appears anxious that the computational demands of AI will soon surpass our terrestrial capabilities. Scattered throughout the SEC filing are allusions to “terawatt-scale annual AI compute growth,” requiring proportional power generation. This is staggering, given that global data centers currently consume around 40 gigawatts.

This underscores Musk’s “first principles” approach. He has likely concluded that the world will need an additional terawatt of computing power each year and has reversed-engineered his hypotheses from that assumption. “We believe that third-party estimates on data center demand are limited by practical supply constraints that exist in a terrestrial context and that the power shortfall may be much larger than current research suggests,” the company claims.

Is this doable? One might argue affirmatively. However, it’s essential to recognize that humanity currently consumes around 35,000 terawatt-hours of energy annually, equivalent to roughly 4 terawatts consistently. Energy demands have surged recently, and while AI is likely undergoing rapid growth, it could either keep climbing or plateau. This remains uncertain, but Musk excels at identifying trends at inflection points and predicting from there.

Here is where Musk’s challenges return to Earth. While I may not possess a rocket scientist’s qualifications, transporting solar panels on a flatbed truck likely consumes less energy than launching them into orbit. Additionally, solar panels designed for outer space will need to be manufactured at an enormous scale. These challenges aren’t insurmountable but could serve as distractions. We have only just begun to tap into solar’s potential here on Earth.

The ideal should not hinder the achievable. There are immense opportunities to improve conditions on Earth while also pursuing our aspirations among the stars.

Just three years ago, Musk and his team at Tesla introduced “Master Plan Part 3,” carefully detailing a “plan to eliminate fossil fuels.” A practical starting point might be found within xAI’s data centers.