Orbital Plans to Launch AI Data Centers into Space to Solve Power and Cooling Issues

The company is developing AI data centers designed to operate in low Earth orbit, powered by solar energy and cooled by radiating heat directly into space.

Apr. 13, 2026 at 7:05pm

A highly detailed, glowing 3D illustration of a network of interconnected satellites in low Earth orbit, featuring recognizable components like solar panels and computing modules, conveying the concept of space-based AI infrastructure.Orbital's space-based AI data centers aim to bypass the power and cooling constraints of terrestrial facilities, harnessing the abundant solar energy and natural cooling of the space environment.Los Angeles Today

Orbital, a startup backed by a16z Speedrun, is planning to launch AI data centers into space to solve the growing power and cooling challenges facing terrestrial data centers. The company is developing satellites that will host Nvidia-powered compute hardware and run AI inference workloads, taking advantage of the continuous power generation and natural cooling provided by the vacuum of space.

Why it matters

The race to build ever larger AI models has created an unexpected bottleneck, with power consumption and cooling costs becoming a major constraint on the pace of AI development. Orbital believes that by moving AI infrastructure into low Earth orbit, they can bypass these terrestrial limitations and enable continued advancements in AI technology.

The details

Orbital is preparing its first test mission, Orbital-1, scheduled to launch on a SpaceX Falcon 9 in April 2027. The satellite will validate sustained GPU operation in orbit, test radiation resilience, and begin running AI inference workloads. The company chose inference over training because inference tasks can run independently across a satellite network, making them better suited to a distributed satellite constellation.

  • Orbital-1 is scheduled to launch on a SpaceX Falcon 9 in April 2027.
  • Orbital is currently standing up Factory-1, a robotic satellite assembly facility in Los Angeles.

The players

Orbital

A startup developing AI data centers designed to operate in low Earth orbit, powered by solar energy and cooled by radiating heat directly into space.

Euwyn Poon

The founder of Orbital.

a16z Speedrun

The venture capital firm that is backing Orbital's funding.

SpaceX

The company that will launch Orbital's Orbital-1 test mission on a Falcon 9 rocket.

Nvidia

The company providing the compute hardware that will be hosted on Orbital's satellites.

Got photos? Submit your photos here. ›

What they’re saying

“Launch costs are going to collapse with Starship, going from $7,000/kg on the Falcon 9 today towards the target of $10/kg.”

— Euwyn Poon, Founder, Orbital

“In orbit, the marginal cost of energy is zero. The sun delivers 1,361 watts per square meter in LEO, constantly, with no fuel cost, no grid fees, no utility contract.”

— Euwyn Poon, Founder, Orbital

What’s next

Orbital is preparing to launch its Orbital-1 test mission in April 2027 to validate the company's technology and begin running AI inference workloads from space.

The takeaway

By moving AI infrastructure into low Earth orbit, Orbital aims to bypass the growing power and cooling constraints facing terrestrial data centers, enabling continued advancements in AI technology. This innovative approach could reshape the future of AI development and infrastructure.