"Data Centers in Space"
Starcloud (founded 2024, Redmond WA) launched Starcloud-1, the first satellite with an Nvidia H100 GPU, targeting gigawatt-scale orbital AI data centers.
Starcloud, Inc. is a space-based AI computing company headquartered in Redmond, Washington, founded in January 2024 in El Segundo, California under the name Lumen Orbit before relocating to Redmond in February 2024 and later rebranding as Starcloud. The company was founded by CEO Philip Johnston (formerly of McKinsey & Company), Chief Engineer Adi Oltean (formerly of SpaceX and Microsoft Azure), and CTO Ezra Feilden (formerly of Airbus Defence and Space). Starcloud's core proposition is building orbital data centers that use continuous solar power and radiative cooling in space to run GPU-based AI compute, with the company stating this approach could reduce electricity costs by roughly 90% compared to terrestrial data centers while avoiding land-based permitting constraints, ultimately aiming to scale to gigawatt-level compute capacity in orbit. In November 2025, Starcloud launched its first satellite, Starcloud-1, carrying an Nvidia H100 GPU that the company describes as delivering roughly 100 times more powerful AI compute than had previously been placed in orbit. Following launch, Starcloud reported running versions of Google's Gemini and Gemma language models aboard the satellite and conducting in-orbit training of Andrej Karpathy's nanoGPT model, which the company characterizes as the first large language model training performed in space. In February 2026, Starcloud filed a proposal with the U.S. Federal Communications Commission for a constellation of up to 88,000 satellites as part of its longer-term orbital data-center roadmap, which includes planned follow-on satellites Starcloud-2 through Starcloud-4. The company has raised approximately $34 million in seed funding by March 2025 and a $170 million Series A round in March 2026 that valued the company at roughly $1.1 billion, with investors including Sequoia Capital, Andreessen Horowitz, and In-Q-Tel, and it has also received backing and GPU support from Nvidia. As of 2026 the company employed approximately 15 people, reflecting its early operating stage. Public reporting from mid-2026 has also noted that scaling orbital GPU operations beyond a single satellite remains an operational challenge, underscoring that space-based AI compute is still an early-stage, unproven infrastructure category rather than a mature commercial service. Starcloud has not published third-party certification information, and no official Facebook or YouTube company pages were identified during this review.
Primary type & automation activities this supplier delivers:
Contact Starcloud
WEBSITE
https://www.starcloud.com/HEADQUARTERS
United States
Company Facts
Founded
2024
Primary Role
Manufacturer
Company Size
employees 10-50
Primary Region
North America
Annual Sales
-
Funding Stage
Early-Stage VC
Funding Total
-
Keywords
At the 2026 World Artificial Intelligence Conference, Shanghai Electric presented its advanced embodied intelligence solutions for industrial applications. The company emphasized its capabilities in embodied AI robots, core components, and AI-native smart factory technologies, showcasing a comprehensive portfolio designed for various industrial scenarios. This development is significant as it highlights the integration of machine precision with human expertise, aiming to enhance collaboration in manufacturing processes. Shanghai Electric's robotics solutions address critical tasks such as connector insertion, intelligent assembly, and flexible sorting, showcasing innovations like the planetary roller screw and DexHand dexterous hand. Looking ahead, Shanghai Electric's launch of 51 AI models under the “StarCloud Intelligent Manufacturing” series marks a shift towards data-driven operations in manufacturing. The release of the “AI-Native Smart Factory Technology White Paper” outlines a new architecture for real-time production optimization, indicating a transformative approach to industrial automation. No further timeline was disclosed at the time of publication.
RoboticsAndAutomationNews.com Jul 30, 2026 Artificial Intelligence Robotics AI-native smart factory embodied ai humanoid robots industrial aiAt the World Economic Forum in Davos this January, SpaceX founder Elon Musk announced plans to establish orbital data centers in space, predicting that they will become the most cost-effective solution for artificial intelligence (AI) within two to three years. Following this declaration, SpaceX submitted an application to the Federal Communications Commission for a constellation of up to 1 million satellites in low Earth orbit, aimed at supporting this ambitious project. However, experts caution that the logistics of deploying such a vast number of satellites are daunting. Currently, there are approximately 14,500 active satellites in orbit, with SpaceX's Starlink making up two-thirds of that total. To launch 1 million satellites, SpaceX would need to conduct over 16,000 dedicated launches, a feat that could take decades given current launch capacities. Challenges also extend to the technical feasibility of cooling advanced computing hardware in space, as highlighted by the difficulties faced by startups like Starcloud, which has struggled to operate even a single GPU in orbit. Concerns have been raised about the potential for increased space debris and the impact on astronomical observations. Despite these hurdles, analysts suggest that the push for orbital data centers is driven by the growing demand for AI computing power and the need for energy-efficient solutions. While Musk's timelines may be optimistic, industry experts believe that the concept is gaining traction, with major players beginning to invest in the necessary infrastructure to explore the viability of space-based data centers.
IEEESpectrumAI Jul 01, 2026 Orbital-data-centers Satellites Spacex Elon-musk Starcloud AiFrom Starcloud's $170M Series A to Google's Project Suncatcher and China's operational LEO constellations, AI space data centers are moving from concept to infrastructure. Here's what the robotics industry needs to know.
ByKelly Stone Apr 03, 2026