About
Starcloud is building orbital data centers that initially provide GPU compute to other satellites and eventually support large AI workloads in space. Its initial committed customers include Earth-observation and DOW spacecraft, while its differentiation is combining abundant solar energy, radiative cooling, and unconstrained orbital scaling to address terrestrial power and permitting constraints.
Market
Starcloud competes in the emerging space-based data-center and orbital AI-infrastructure market, positioning itself as an infrastructure provider for GPU compute rather than simply a satellite or launch company. Its differentiation is a first-mover operational demonstration with an NVIDIA H100 and in-orbit LLM training, combined with a thesis that continuous solar power, radiative cooling, and unconstrained orbital expansion can lower energy costs and support gigawatt-scale AI clusters. Competitors pursue different models: Google is developing TPU-based Project Suncatcher, SpaceX is proposing a very large satellite constellation, Axiom is deploying orbital nodes connected to the ISS, and Lonestar is focused more on lunar data storage and disaster recovery.
Starcloud initially targets satellite operators and Earth-observation providers that need on-orbit GPU compute and low-latency inference, including SAR-data users. Its longer-term customers are hyperscalers, cloud providers, and AI companies operating large training clusters and seeking capacity that can scale from megawatts to gigawatts.
At a Glance
Problem
Starcloud is addressing the energy, cooling, and deployment constraints of terrestrial data centers, especially for energy-intensive AI compute. Its thesis is that orbital facilities can use near-continuous solar power without batteries and radiative cooling, while avoiding restrictive permitting and scaling toward gigawatt capacity. The company advertises up to 90% lower electricity costs, though that figure is a company claim rather than independently validated evidence.
The primary use case is large-scale AI infrastructure: Starcloud describes its ambition as deploying the largest training clusters in space to support the future of AI. A nearer-term use case is providing cloud compute for other satellites, which could process data in orbit rather than transmitting all of it back to Earth.
Product / Service
Starcloud is building data centers in space and plans to deliver compute as a cloud service, initially for other satellite operators. Its first demonstrated platform, Starcloud-1, launched in November 2025 with an NVIDIA H100 GPU and reportedly trained an LLM in orbit, providing an early proof point for orbital AI computing rather than merely a terrestrial data-center concept.
The longer-term architecture combines large solar arrays and radiative-cooling systems with expandable orbital infrastructure. The intended benefit is continuous solar-powered operation, lower electricity costs, rapid scaling without terrestrial permitting constraints, and access to compute close to space-generated data. The company’s roadmap points toward much larger orbital clusters, while Starcloud-1 establishes the initial hardware and software model.
Market
Starcloud competes in the emerging market for space-based or orbital data centers, combining cloud computing, satellite infrastructure, and AI hardware. The competitive field includes other startups such as Axiom Space and Lonestar, as well as larger technology and space companies including Google and SpaceX, which are also pursuing orbital data-center concepts.
The company appears to be in an early commercialization phase rather than a mature revenue business. Its traction includes the Starcloud-1 orbital H100 and LLM milestone, a reported $170 million Series A at a $1.1 billion valuation in March 2026, and approximately $200 million in total funding. Available reporting said Starcloud had no disclosed commercial revenue as of April 2026, so the strongest current signals are technical validation and investor backing, not demonstrated recurring sales.
Founders & Leadership
Funding History
Nebular, Caffeinated Capital, Plug & Play, Everywhere Ventures, Tiny.vc, Sterling Road, Pareto Holdings, Foreword Ventures, Sequoia Scout Fund participants
NFX
Not publicly identified
Benchmark, EQT Ventures
Recent News
Starcloud announced a contract for more than 50 Starlink Mini Lasers, enough to equip at least 25 satellites with two lasers each for its planned orbital data-center network.
SpaceNews reported that Starcloud was seeking at least $200 million in additional funding, potentially doubling the company's valuation to approximately $2.2 billion.
Starcloud raised $170 million to accelerate development of its next-generation spacecraft, reaching a reported valuation of $1.1 billion.
Data Center Dynamics reported that Starcloud was running an AI model in space, demonstrating an early orbital-computing use case ahead of the company's broader space data-center ambitions.
CNBC reported that Starcloud had trained its first AI model in space, highlighting the company's progress with orbital computing and the emerging competition around space-based data centers.
Starcloud-1 successfully launched and separated from its SpaceX rocket, putting an NVIDIA H100 GPU into orbit as a trial deployment for Starcloud's space data-center strategy.
Crusoe partnered with Starcloud to deploy Crusoe Cloud on a Starcloud satellite scheduled for launch in late 2026. The arrangement is intended to provide limited GPU capacity in orbit.
NVIDIA described Starcloud's planned November satellite launch, which was expected to mark the NVIDIA H100 GPU's debut in space. The company has also outlined a long-term vision for a 5-gigawatt orbital data center.
Active Roles
12Business Model
Starcloud is designed to monetize orbital infrastructure by selling processing power and cloud or edge services to other spacecraft, particularly Earth-observation and DOW satellites. It is also pursuing energy offtake agreements with hyperscalers; public sources do not disclose a finalized pricing schedule.