As global investment surges into hyperscale AI infrastructure—from the landmark 500 MW xAI–HUMAIN–NVIDIA initiative to trillion-dollar national AI strategies—one reality is becoming clear: the global race for compute is accelerating faster than terrestrial infrastructure can support.
Today, 38 Labs is proud to announce 38 Space, our pioneering initiative to build AI-native orbital data centers— a transformative leap that redefines how and where advanced AI systems will be trained in the coming decade.
While nations and corporations compete to build ever-larger data centers on Earth, these facilities face unavoidable constraints: energy-intensive cooling demands, land and environmental permitting, grid limitations, water scarcity, and growing community pushback. Even the newest mega-projects, such as Elon Musk’s 500 MW xAI data center in Saudi Arabia, highlight the immense capital, regulatory, and environmental weight required to scale compute on the ground.
38 Space takes a fundamentally different path.
Instead of scaling compute within Earth’s limitations, we are extending AI infrastructure beyond them.
Why Space?
Space offers what Earth cannot:
- Near-continuous solar power (≈99% sunlight in GEO)
- Passive radiant cooling using the vacuum of space
- Zero land, water, and emissions footprint
- Orbital safety from natural disasters or geopolitical risk
- Global coverage and real-time access to satellite data streams
In low-Earth orbit (LEO) and geosynchronous orbit (GEO), AI infrastructure becomes:
- Scalable — through modular, launch-ready compute nodes
- Sustainable — powered purely by the Sun
- High-performance — cooled by physics, not megawatt chillers
- Globally accessible — forming a computation layer around the planet
Introducing 38 Space: AI-Native Infrastructure in Orbit
Unlike traditional “data centers in space” concepts, 38 Space is designed from the ground up for AI, not lifted from Earth. Our system integrates:
- Orbital compute nodes engineered for high-density AI hardware
- Ultra-efficient solar arrays and thermal radiators
- A redundant inter-satellite mesh for distributed AI training
- Automated in-orbit maintenance, upgrades, and scaling
This is AI infrastructure unconstrained by geography, unlocked by physics, and deployed with the cadence of reusable launch systems.
Building the Next Compute Frontier
As global demand for AI compute continues its exponential climb—driven by frontier models, autonomous systems, and nation-scale digital transformation—the world needs new paradigms, not larger buildings.
38 Space is our answer to that need.
We are positioning 38 Labs at the forefront of a new era: one where AI is trained above the clouds, powered by the Sun, cooled by the vacuum of space, and interconnected across orbital swarms.
The race for AI supremacy is no longer just terrestrial. It is orbital.
With 38 Space, we are stepping boldly into that future.
