Remember Elon Musk's plan to put AI data centers in space? It appears companies are taking the idea seriously, and one of those companies is Google. According to a new report in the Wall Street Journal, Google is currently in talks with Musk's space exploration company, SpaceX, to strike a deal to launch rockets into space with the intent of putting data centers into orbit.
When SpaceX acquired xAI, Musk's AI company, earlier this year, Musk penned a statement explaining why he decided to combine his companies. One big reason? Data centers in space. While this was not the only reason, it was a main focal point for SpaceX's acquisition of xAI. SpaceX had recently filed with the Federal Communications Commission (FCC) seeking permission to launch "a million satellites" to put AI data centers into orbit.
"Current advances in AI are dependent on large terrestrial data centers, which require immense amounts of power and cooling," Musk explained at the time of the acquisition. "Global electricity demand for AI simply cannot be met with terrestrial solutions, even in the near term, without imposing hardship on communities and the environment. In the long term, space-based AI is obviously the only way to scale."
Google seems to agree with Musk. Late last year, Google announced Project Suncatcher, an initiative to launch prototype satellites by 2027 in order to "one day scale machine learning computer in space." Then, in February, just weeks after SpaceX's acquisition of xAI, Google CEO Sundar Pichai shared that the company was looking into its own orbital data centers. While speaking at the AI Impact Summit in New Delhi, India, Pichai recounted how when growing up in India, he never imagined he'd "one day be spending time with teams figuring out how to put data centers into space."
While Google is still exploring rocket launch options with other companies, the search giant wouldn't be the first to partner with SpaceX in hopes of putting new AI data centers in space. Last week, Anthropic and SpaceX announced a partnership to utilize xAI's data centers in Memphis, Tennessee. The deal also includes future space development as well. A deal with Google would also be extremely beneficial to SpaceX right now, as the company plans its $1.75 trillion IPO in the coming months.
The Growing Need for Space-Based AI Computing
The push for orbital data centers stems from the exponential growth of artificial intelligence workloads. Training large language models like GPT-4 or Gemini requires massive clusters of specialized processors, often consuming tens of megawatts per facility. As AI becomes more pervasive, the energy footprint of data centers is expected to skyrocket. According to the International Energy Agency, global data center electricity consumption could double by 2026, with AI being a major driver. Terrestrial data centers already account for about 1-2% of global electricity use, a figure that is projected to rise rapidly.
Beyond energy, cooling is a critical bottleneck. High-performance chips generate enormous heat, requiring elaborate cooling systems that consume additional water and electricity. In space, however, ambient temperatures are near absolute zero, offering natural heatsinking. Solar power is also abundant in orbit, with no atmosphere to attenuate sunlight. Orbital data centers could potentially be powered by large solar arrays, providing a constant energy source without competing with terrestrial grids.
Space-based computing also offers latency advantages for certain applications. For instance, satellite sensors generating petabytes of data daily could process that data on orbit rather than transmitting it to Earth. This reduces bandwidth requirements and enables real-time analytics for Earth observation, climate monitoring, and telecommunications. SpaceX's Starlink constellation already provides low-latency internet, and integrating AI compute nodes could enhance its capabilities.
Project Suncatcher: Google's Ambition
Google's Project Suncatcher, revealed in 2024, aims to test the feasibility of operating machine learning computers in space. The initiative involves developing small-scale prototype satellites equipped with custom AI accelerators. These satellites would be designed to handle inference tasks for applications like image recognition, natural language processing, and autonomous navigation. The 2027 timeline suggests Google is moving beyond theoretical studies into practical engineering.
One key challenge is radiation. Space is filled with high-energy particles that can disrupt electronics and cause errors in computations. Google's hardware would need to be hardened or use error-correction techniques to ensure reliability. Additionally, thermal management in microgravity requires innovative cooling solutions, such as phase-change materials or heat pipes coupled with radiators.
Another hurdle is connectivity. Orbital data centers must maintain high-bandwidth links to Earth to receive queries and return results. Laser communication systems, already being tested by SpaceX and others, offer gigabit-per-second data rates but require precise pointing and tracking. Google could leverage its existing fiber network and cloud infrastructure to integrate space-based nodes seamlessly.
The economic rationale for Project Suncatcher is also compelling. Launching hardware into space remains expensive, but SpaceX's reusable rockets have dramatically reduced costs. A Falcon 9 launch costs around $67 million, while Starship could lower it further to under $10 million per launch in the future. If Google can achieve a density of computing per kilogram similar to terrestrial racks, the total cost of ownership might become competitive for specialized workloads.
SpaceX's FCC Filing and the xAI Acquisition
SpaceX's FCC filing requesting authorization for "a million satellites" underscores the scale of Musk's vision. The company already operates the largest satellite constellation with Starlink, but increasing it by orders of magnitude would require new orbital slots and spectrum rights. The filing likely anticipates a future where thousands of satellites host AI compute nodes, forming a distributed mesh network in low Earth orbit.
The acquisition of xAI in early 2025 was a strategic move to combine Musk's AI expertise with his space ambitions. xAI had been developing large language models and AI infrastructure, and its team now works alongside SpaceX engineers to design orbital compute systems. Musk has long argued that AI safety and scalability require decentralized computing resources, and space presents an opportunity to create a globally accessible, resilient AI platform.
Anthropic's partnership with SpaceX for using xAI's Memphis data centers and future space development indicates that other AI players are also interested in the space-based model. Anthropic, known for its Claude models, may use the terrestrial data centers first, but the agreement includes provisions for orbital expansion. This signals that the industry sees space as a logical next frontier for AI computing.
Implications for SpaceX's IPO
SpaceX is reportedly planning an IPO with a valuation of $1.75 trillion, which would make it one of the largest public offerings in history. Revenue from launching satellites for Google could provide a significant boost. SpaceX already earns billions from Starlink subscriptions and launch contracts, but a long-term deal with Google would add a high-profile customer and demonstrate the viability of its heavy-lift capabilities for non-communication payloads.
Furthermore, the IPO market is sensitive to narratives about future growth. Positioning SpaceX as the enabler of space-based AI computing could justify its massive valuation. Investors are keen on AI-related opportunities, and SpaceX's involvement in this sector could attract capital from tech-focused funds. If the Google deal materializes, it would validate Musk's vision and likely increase demand for SpaceX shares.
However, there are risks. Regulatory hurdles with the FCC and international space agencies could delay deployment. Environmental concerns about space debris and the impact of large satellite constellations on astronomy also persist. Google and SpaceX would need to address these issues transparently to maintain public and investor confidence.
In summary, the potential collaboration between Google and SpaceX to launch orbital data centers represents a convergence of AI, space, and energy challenges. Both companies are pioneering solutions that could transform how computing is done, moving beyond terrestrial constraints. With Project Suncatcher, SpaceX's FCC filing, and the xAI acquisition, the pieces are falling into place. Whether these plans reach orbit by the end of the decade remains to be seen, but the dialogue between two of the most innovative companies signals a new era for data infrastructure.
Source: Mashable News