Tesla's CEO is highly optimistic about the prospect of computing in space, as Musk believes Earth's energy constraints hinder the advancement of AI.
Tesla's CEO Claims Energy Is a Problem On Earth, But in Outer Space, Chips Will Be the Next Constraint
As the world of AI infrastructure advances, many experts argue that the 'dot-com' moment for this buildout will come when hyperscalers realize there isn't enough energy on the grid to sustain datacenter development, which would ultimately lead to a compute glut. Energy is a massive constraint that the AI bandwagon currently witnesses, given that it is argued that ultimately, the power generation on our planet will hit a "roadblock", which would dramatically increase training costs, making it difficult for hyperscalers to operate. To solve this, Elon suggests having data centers that rotate in space.
In 36 months, but probably closer to 30 months, the most economically compelling place to put AI will be space. The limiting factor once you can get to space is chips, but the limiting factor before you can get to space is power.
- Elon Musk via Dwarkesh Patel
Musk argues that the current power dynamics on Earth in terms of generation and distribution into the grid aren't enough at all for the AI buildout, saying that the US alone consumes just half a terawatt of electricity, but given the pace of the infrastructure progression, deploying datacenters and building plants on the ground isn't a viable option, according to him. The IEA says that within the next four years, data centers alone will see electricity consumption rise by up to 15%, and by 2030, DC could take up 12% of America's total power generation.
All of the United States currently uses only half a terawatt on average. So if you say a terawatt, that would be twice as much electricity as the United States currently consumes. So that’s quite a lot. Can you imagine building that many data centers? That many power plants?
Here's an interesting twist: Musk argues that once the mechanisms are in place for orbital datacenters through Starship for transportation, deployment, and Starlink for networking, chips will become the major bottleneck. Tesla's CEO says they are working with every fab out there, including TSMC's Arizona and Taiwan operations, as well as Samsung's Korean and Texas fabs. Still, he believes that the capacity and delivery times they provide aren't sufficient, which leads to the idea of creating TeraFab.
Well, energy is a huge bottleneck for the current datacenter expansion and the "AI regime", and Tesla's CEO believes that orbital datacenters will solve this problem. Funny enough, Musk claims that this venture would be a side quest towards his plans to reach Mars, saying it would open up another prospect for SpaceX. The idea of data centers in space isn't entirely new, and, interestingly, the startup Starcloud has already deployed NVIDIA's H100 outside Earth. But when you talk about mega-GW buildout in space, that is when things become interesting and questionable.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
