- Momentum is growing behind orbital data centers
- Starcloud just raised millions at a $1.1 billion valuation, helping it compete with SpaceX and Google's Project Suncatcher
- Key questions around the economic viability of such plans remain
Starcloud just crossed a huge milestone, raising $170 million in funding at a $1.1 billion valuation. While $1 billion isn’t exactly a huge sum in the data center world, it is certainly a lot for a startup – and one looking to deploy data centers in space at that.
Starcloud’s monster valuation is reflective of growing enthusiasm around the potential for a new kind of data center: the kind in orbit. The theory goes that space-based data centers can sidestep key terrestrial constraints tied to the availability of land and – more importantly – power. After all, what is the sun if not a massive source of energy? And the company is just one of several big names looking to the stars.
Here are the names you need to know in this space:
Starcloud
Starcloud, formerly known as Lumen Orbit before rebranding in early 2025, has raised $200 million to date. In November 2025, it launched its Starcloud-1 satellite, equipped with a Nvidia H100 PGU, into orbit. The following month it began running Google’s Gemini model on the chip and also used it to train Andrej Karpathy’s nano-GPT LLM.
One of the headline features of Starcloud’s pitch is that its satellite data centers will be able to operate with 10x lower energy costs than terrestrial data centers thanks to the use of sun-synchronous orbits that can maximize solar power generation.
The company has plans to launch three additional satellites. These include Starcloud-2, which will feature a complete GPU cluster, persistent storage and thermal and power systems in a smaller form factor. Connectivity to earth will be provided by an optical link, with a third party supplying an RF-based backhaul network. The goal is to have this satellite launched in a sun-synchronous orbit and operational by 2027.
SpaceX
Elon Musk’s SpaceX has perhaps the most ambitious – and outlandish – plans for orbital data centers. The company is seeking approval to launch a constellation of 1 million satellites, which would deliver 100 kW of compute power per tonne. These would be launched in a sun-synchronous orbit to maximize solar power generation.
“The constant power generation in sun-synchronous orbits will reduce the need to rely on batteries and power cycling, making these orbits ideal for energy-intensive AI workloads that demand constant, reliable compute capacity,” the company wrote in its application.
A target launch date has not yet been disclosed.
Project Suncatcher
Announced in November 2025, Project Suncatcher is Google’s moonshot project aimed at launching solar-powered orbital data centers equipped with the company’s Tensor Processing Units (TPUs). Its first step will be launching “two prototype satellites by early 2027” to see if its chips can withstand the sun’s radiation.
Like Starcloud and SpaceX, Google is looking to sun-synchronous orbits for its birds. Its rationale is the same as SpaceX’s: “In the right orbit, a solar panel can be up to 8 times more productive than on earth, and produce power nearly continuously, reducing the need for batteries. In the future, space may be the best place to scale AI compute” it noted in a blog.
It’s not clear how many satellites Google is planning. Its blog mentions “compact” constellations and uses an 81-satellite configuration to illustrate its plans. But beyond that, it didn’t mention the scale of its ambitions.
Other players
Others are also looking to space as a new home for data centers. Axiom Space, for instance, last year launched its first data center prototype unit (which runs on Red Hat Device Edge) aboard the International Space Station in August. The unit was able to run cloud computing, AI, data fusion and space cybersecurity workloads, and was used to test applications and demonstrate initial capabilities.
In September, it announced a collaboration to deploy orbital data center infrastructure to the ISS in 2027. The company was awarded up to $5.5 million from the Texas Space Commission to support its orbital data center plans. In January 2026, it launched its first two data center nodes into low earth orbit.
Additionally, Lonestar Data Holdings is pursuing data center deployments on the moon, though its target use case is more disaster recovery than edge processing.
Why it matters and key questions
Fierce has extensively covered issues around electrical power consumption and how grid constraints are shaping interest in future power generation and data center siting.
It’s clear that the allure of an endless supply of solar power is a large part of what’s driving interest in orbital data centers. But questions around the economic viability of and launch constraints around these plans remain.
Andrew McCalip, an aerospace engineer and head of Research and Development at Varda Space Industries, did the math on the economic side of things.
He argued that beyond just “energy is cheaper” the goal of these kinds of initiatives is a lower cost per watt of usable power for computing, one that beats the terrestrial number. “Orbit has to win on cost, or it has to admit it's doing something else entirely,” he wrote.
Right now, according to McCalip’s model, orbital deployments have substantial cost gap to overcome. His model – which didn’t account for the required compute but everything upstream of that – found that the cost of 1 GW of orbital solar compute would cost $51.1 billion versus $15.9 billion for the same on Earth.
But Google seems to think the cost equation could shift in orbital’s favor.
Though it acknowledged historically high launch costs, Google argued in a blog that if launch costs come down at a sustained rate through the mid-2030s, “the cost of launching and operating a space-based data center could become roughly comparable to the reported energy costs of an equivalent terrestrial data center on a per-kilowatt/year basis.”
Launch costs, though, are just one piece of this incredibly complex puzzle. Perhaps McCalip’s most important takeaway is that the success of orbital ventures will rely heavily on control. And control means vertical integration.
“If you have to buy launch, buy buses, buy power hardware, buy deployment, and pay margin at every interface, you never get there. The margin stack and the mass tax eat you alive. Vertical integration isn't a nice-to-have. It's the whole ballgame,” he concluded.