"Space Data Center" becomes a battleground for AI? Musk and Bezos are at odds, and Altman also wants to get involved

The trillion-dollar data center boom in artificial intelligence (AI) is giving rise to a whole new battleground: outer space. Tech billionaires Elon Musk and Jeff Bezos, after a long-standing competition in the fields of rockets and satellites, are extending their race to orbital AI data centers, attempting to move massive computing infrastructure off Earth.

On December 10, The Wall Street Journal reported, citing informed sources, that Musk's SpaceX plans to use an upgraded version of its Starlink satellites to carry AI computing loads and is promoting this technology as part of its stock offering, a plan that could value SpaceX at $800 billion. Meanwhile, an insider stated that Bezos's Blue Origin has also formed a team that has been conducting research and development for over a year on the technology needed for orbital AI data centers.

The participants in this race are far more than just these two giants. Reports indicate that OpenAI's CEO Sam Altman has also explored the possibility of acquiring a rocket operator to deploy AI computing capabilities in space. Tech giants like Google are also actively positioning themselves in this emerging trend, which signals a potential intersection of the AI and aerospace industries, driven primarily by the need to address the growing energy consumption challenges of AI on Earth.

While supporters are excited about the prospect of harnessing unlimited solar energy from space, sending data centers into orbit faces significant engineering hurdles and cost-effectiveness questions. Skeptics argue that the technological risks are underestimated and that it will be difficult to compete with ground facilities in the short term. For investors, this represents both an imaginative future opportunity and a high degree of uncertainty.

Earth "cannot accommodate" AI ambitions, space becomes a new energy solution

The core logic of sending data centers into space is to escape the physical limitations on Earth, especially the astonishing power required for AI model training and inference. Supporters envision a future where satellites loaded with AI chips populate orbit, operating on a continuous supply of solar energy and sending processed data back to Earth.

"The idea of moving resource-intensive infrastructure off Earth has been around for years, but it requires a decrease in launch and satellite costs. We are approaching that critical point," said Will Marshall, CEO of satellite operator Planet Labs.

Veteran tech investor Gavin Baker also believes that, from first principles, space data centers have energy and cooling advantages that ground facilities cannot match. In space, satellites can receive solar energy that is 30% stronger and up to six times more abundant than on the Earth's surface, without the need for expensive energy storage batteries. Additionally, using the near absolute zero environment of space for cooling is almost "free," which can eliminate one of the most complex and costly cooling systems of ground data centers. Baker predicts that space data centers could become one of the most significant technological breakthroughs in the next three to four years.

Giants Compete: Musk's "Starship" vs. Bezos's "New Glenn"

In this space race, launch capability is the key to success or failure. Musk and Bezos are both leveraging their aerospace companies' next-generation heavy-lift rockets to pave the way for grand visions of space data centers.

According to insiders, SpaceX's AI computing technology will be installed on upgraded satellites designed specifically for the Starship. The Starship is a giant, fully reusable rocket being developed by SpaceX, aimed at significantly reducing launch costs. Musk stated last month on social media platform X that the Starship should be able to launch approximately 300 gigawatts (GW) of solar AI satellites into orbit each year, potentially reaching 500 GW.

On the other hand, Bezos's Blue Origin has made significant progress this year on its long-developed New Glenn rocket project. This partially reusable rocket features a massive fairing designed to launch a large number of satellites into orbit in a single mission. Bezos mentioned at an event in October that moving data centers to orbit makes sense and predicted that its costs would surpass ground-based AI infrastructure within 20 years or less.

"This ultimately comes down to launch (capability)," said Jonny Dyer, CEO of satellite company Muon Space. For investors, this means that companies with reliable, low-cost, high-capacity launch tools will hold a significant advantage in this emerging competition.

New Players Entering the Field: Altman and Google Also Making Moves

In addition to Musk and Bezos, other tech leaders have shown strong interest in orbital data centers, trying to carve out a share of this potentially massive market.

According to The Wall Street Journal, OpenAI CEO Sam Altman has explored the possibility of having his company take over a rocket operator to leverage its vehicles for deploying AI computing capabilities in space. He believes that the enormous demand for computing power from AI systems may ultimately require so much electricity that space becomes a better option.

Meanwhile, Google has reached an agreement with Planet Labs to deploy two test satellites carrying Google AI chips (Tensor Processing Units) into orbit by early 2027. Google describes this project as one of its "moonshot plans," highlighting the challenges of deploying a network of satellite data centers on a large scale. According to Travis Beals, an executive at Google overseeing the project, assuming the use of 100-kilowatt satellites, 10,000 satellites would be needed to replicate the computing power of a one-gigawatt ground data center.

Additionally, former Google CEO Eric Schmidt, who now heads rocket company Relativity Space, has also discussed orbital data centers. IBM's Red Hat software business and Axiom Space launched a data computing prototype in August this year. Venture-backed startups like Aetherflux and Starcloud are also making plans to compete with large enterprises.

Real Challenges: Cost, Technology, and Scaling Issues Remain Unresolved

Despite the enticing prospects, transitioning space data centers from concept to reality still faces numerous obstacles.

First are the severe technical challenges. AI chips operating in orbit need to address a series of issues, including temperature management, cosmic radiation protection, and how to transmit massive amounts of data back to Earth without noticeable delays. Any oversight in any of these areas could lead to a complete system failure.

Secondly, there are costs and scalability. Skeptics argue that supporters underestimate the technological risks and question whether space-based data centers can compete with ground facilities in terms of cost, especially as ground power and other limiting factors are alleviated. As pointed out by Google executives, achieving a scale comparable to ground operations would require launching thousands of satellites, which places extremely high demands on current launch capabilities and cost control.

Investor Gavin Baker also noted that while the application prospects for space data centers are clearer for inference tasks, fully migrating AI model training tasks to space may take longer due to the large scale involved. The road to space AI computing power is destined to be long and arduous