Nvidia CEO Jensen Huang said the company is capable of deploying data centers in space and is exploring how to scale that capability.
This push toward orbital infrastructure represents a strategic attempt to expand AI compute capacity beyond the physical and environmental constraints of Earth. As demand for artificial intelligence grows, the ability to process data in space could reduce latency for satellite networks and unlock new frontiers for high-performance computing.
Huang discussed the concept during a CNBC Television interview. He said that while the technical ability to place data centers in orbit is established, the primary hurdle is the volume of deployment. "I think the ability to deploy data centers in space is a known ability," Huang said. "The question is, how do we scale that up?"
Despite the technical feasibility, the financial viability of such projects remains a significant barrier. Huang said the economics of space data centers are poor right now, but will improve.
This vision for space-based AI coincides with broader industry growth and high financial stakes for hardware providers. Nvidia's sales forecast for AI chips has reached $1 trillion [3]. The company's focus on expanding data center capabilities is central to maintaining this growth trajectory as traditional terrestrial sites face power and cooling limits.
Other firms are already moving toward this orbital goal. In March 2026, a company called Starcloud raised $170 million [4] in Series A funding to build data centers in space. Following that funding round, Starcloud reached a valuation of $1.1 billion [5].
These developments indicate a shift toward a hybrid compute model where AI processing is distributed between ground stations and orbital nodes. Huang's comments suggest that Nvidia views the transition to space as an inevitability once the cost of launch and maintenance decreases.
“The question is, how do we scale that up?”
The move toward space-based data centers suggests that the AI industry is approaching a terrestrial ceiling regarding energy and land use. By shifting compute to orbit, companies can potentially bypass local power grid limitations and create a global, low-latency AI layer. However, the reliance on emerging firms like Starcloud and the current 'poor' economics indicate that this transition will be gradual and dependent on the falling cost of space logistics.
