Google to launch First Space Data Centers in 2027
- Marijan Hassan - Tech Journalist
- 12 minutes ago
- 2 min read
The space race for Artificial Intelligence has officially begun. Google CEO Sundar Pichai has confirmed that the company is moving forward with its ambitious plan to launch its first experimental space-based data centers into low-Earth orbit by 2027.

The initiative, internally named Project Suncatcher, is Google's most audacious "moonshot" to date. It is framed as a long-term, necessary solution to meet the rapidly escalating, near-insatiable power demands of advanced AI models like Gemini 3, without overwhelming Earth's increasingly strained power grids and water resources.
The vision: AI compute powered by the sun
Pichai, speaking in a recent interview, confirmed the timeline and the ultimate vision:
"One of our moonshots is: How do we one day have data centers in space so that we can better harness the energy from the Sun, which is one hundred trillion times more energy than what we produce on all of Earth today?"
Google's research suggests that by relocating the compute required for massive machine-learning workloads to space, they can leverage several advantages:
Unlimited solar energy: Solar panels in orbit receive constant, direct sunlight, minimizing downtime and making power generation significantly more efficient than on Earth.
Reduced terrestrial strain: The move would alleviate the environmental pressure caused by data centers, which currently consume billions of gallons of water for cooling and contribute significantly to carbon emissions through reliance on terrestrial power grids.
Proof of concept
The initial phase involves launching two prototype satellites in partnership with firms like Planet Labs. These satellites will carry "tiny racks of machines" equipped with Google's custom Tensor Processing Units (TPUs) to test radiation tolerance, thermal management, and reliable operation in space.
Pichai added that within a decade, he expects the construction of off-planet data centers to become a "normal way to build data centers."
The challenges: Engineering and economics
While the promise of Project Suncatcher is immense, Google acknowledges the staggering technical and economic hurdles:
Thermal management: Dealing with waste heat in the vacuum of space without the aind of water or air is an enormous engineering challenge.
Cost and repair: The high cost of launching hardware into orbit, coupled with the inability to send an engineer up for simple repairs, requires systems to be highly resilient and autonomous.
Data transmission: Data computed in space must be sent back to Earth with low latency, requiring complex, high-bandwidth free-space optical links (laser communication) between the orbiting facilities and ground stations.
Despite these obstacles, Google is moving ahead aggressively, betting that the cost of space launches will continue to drop, eventually making space-based data centers cost-competitive with terrestrial ones, and cementing its lead in the AI infrastructure race.













