Starcloud, a U.S.-based startup, has reached a $1.1 billion valuation after raising $170 million in a Series A round. The company builds data centers in orbit. Its goal is simple: move heavy AI computing off Earth.
Starcloud is based in Redmond, Washington. It wants to run AI workloads on satellites instead of land-based data centers. The logic is clear. Space offers steady solar power, cold temperatures, and no land limits.
On Earth, data centers face rising pressure. Power grids strain under demand. Land is scarce in key regions. Cooling systems use large amounts of water and energy. Starcloud aims to bypass all of that.
In orbit, satellites can capture near-constant sunlight. They can also release heat into space more easily than systems on Earth. This setup could lower energy costs over time.
Starcloud has already tested its idea. In late 2025, it launched Starcloud-1 on a mission with SpaceX. The satellite carried a high-end GPU from Nvidia.
The hardware was a Nvidia H100. This chip is widely used for AI training. In orbit, the system trained a small language model and ran inference tasks. It even executed Google’s Gemma model.
This marked the first time an AI model trained in space. It showed that standard AI hardware can work in orbit without major changes.
Starcloud Hits $1.1B Unicorn Status to Build 88,000-Satellite Data Center
The $170 million Series A round values Starcloud at $1.1 billion. That makes it a unicorn just 17 months after its Y Combinator debut. It is the fastest to reach that mark in the accelerator’s history.
Benchmark and EQT Ventures led the round. Other institutional and angel investors joined. Total funding now stands near $200 million.
The speed and size of this round signal strong interest in new AI infrastructure. Investors see limits in current data center models. They are looking for alternatives that can scale.
Starcloud’s vision goes far beyond one satellite. It plans to build a full orbital data center network. The roadmap includes a constellation of up to 88,000 compute satellites.
These satellites would act like nodes in a distributed cloud. Users could run AI workloads in space, then send results back to Earth.
The company also plans to deploy larger platforms. These would resemble full data centers, but in orbit. Over time, Starcloud expects costs to fall as launch prices drop.
Much of this depends on next-generation rockets like Starship. Lower launch costs would make it cheaper to send hardware into orbit at scale.
The next step is Starcloud-2, set for launch in October 2026. This satellite will be far more powerful than the first.
It will offer about 100 times more power capacity. It will also use Nvidia’s newer Blackwell GPU architecture. In addition, it will include hardware from Amazon Web Services.
Starcloud plans to test multiple workloads. These include AI training, cloud services, and even crypto mining. The goal is to see how different use cases perform in space.
How Starcloud and the New Space Race Are Taking AI Beyond Earth
Starcloud is not alone in this idea. The push to move computing off Earth is growing. Companies like Blue Origin and SpaceX are building the launch systems needed to support this shift.
AI demand continues to rise at a fast pace. Training large models requires vast amounts of power and hardware. This creates pressure on national grids and local infrastructure.
Orbital computing offers one possible path forward. It shifts energy use away from cities and industrial zones. It also spreads compute across a global network in space.
The key advantage is energy. Space-based systems can rely on solar power without interruption. They can also cool hardware without complex systems. This could lower operating costs over time.
Scalability is another factor. On Earth, building new data centers takes years. In space, adding more satellites could be faster once launch systems mature.
There are also policy questions. Governments will need to manage orbital traffic, space debris, and resource use. AI infrastructure may soon extend beyond national borders.
Starcloud is still early in its journey. Many challenges remain, especially around cost and reliability. But its early results show that orbital AI computing is possible.
If launch costs fall and hardware keeps improving, space could become a new layer of the cloud. Starcloud is betting that this shift will happen within the next few years.
