TAKING DATA INFRASTRUCTURE OFF THE PLANET

Washington start-up Starcloud to launch world’s first orbital data centre November 2025

“A typical terrestrial data centre can use about 1 million gallons of water daily --equivalent to the daily water consumption of approximately 300,000 people. A data center in Space uses zero fresh water.”

-- Philip Johnston, CEO and Founder Starcloud

Interview with Philip Johnston

By Suzanne Forcese

WT: Philip, please introduce yourself and your co-founders to our viewers and describe the Starcloud raison d’être.

Johnston: I am a second-time founder who previously spent time at McKinsey & Co. working on satellite projects for national space agencies. I have an MPA in National Security & Technology from Harvard University, an MBA from Wharton, an MA in Applied Mathematics & Theoretical Physics from Columbia University, and am a CFA Charter holder.

My co-founder Ezra Feilden has a decade of experience with satellite design, specializing in deployable solar arrays and large deployable structures. Ezra comes from Airbus Defense & Space (SSTL) and Oxford Space Systems, where he worked on many missions, including NASA's Lunar Pathfinder. Ezra has a PhD in Materials Engineering from Imperial College London.

My third co-founder Adi Oltean has a software and hardware background. As part of SpaceX and Microsoft he delivered key features in satellite networks, operating systems, cloud and machine learning infra. His focus areas at Starcloud include software, hardware and engineering design aspects of our satellite constellation.

We came together early in 2024 because of the constraints we are facing in water and energy consumption. AI is growing rapidly but so is the energy demand for its development. Data centres out in orbit would have unlimited access to the power of the sun and the cooling capacities of space.

WT: Can you give us an overview of the Starcloud concept and expected launch dates please.

Johnston: Unlike Earth, where data centres face space and energy limitations, orbital centers can expand up to 50 times larger. These modules, each the size of five shipping containers, combine into massive 4km by 4km solar-powered structures with radiators stretching a kilometer long.

While the initial launch has an environmental cost, over time these space data centers emit ten times less CO2 than Earth-based ones powered by fossil fuel. They also help preserve land by avoiding the need to cover forests and fields with solar panels. In addition to supporting AI, the vacuum and chill of space offer optimal conditions for quantum computing -– eliminating the need for energy-intensive cooling on Earth.

With AI-related electricity consumption expected to grow 50% annually through 2030, Starcould plans to launch our first orbital data centre in November, followed by a more powerful version in 2026.

We believe that most data centres will be in space within a decade.

WT: The idea of most data centres in space withnin 10 years sounds so futuristic. Has interest in this concept reached mainstream?

Johnston: I was actully very surprised to realize that this idea is becoming accepted in certain communities. Very recently Tucson, Arivona voted against data centres in an impassioned plea because of the enormous water usage and proposed building them in space

WT: What can you tell us about the first upcoming launch? And Starcloud-2?

Johnston: Starcloud’s 132-pound (60-kilogram) demonstrator satellite is destined to be launched into low Earth orbit this November by the new SpaceX Falcon 9 rocket (which is 100 times cheaper and faster than any other rocket) as part of the Bandwagon 4 rideshare mission.

It will be assembled in modules. We will be running 100 times more powerful GPU compute than has ever been operated in space, with top-of-the-line, data-center-grade terrestrial NVIDIA GPUs on board.

The demonstrator satellite will help to test training, inference and edge compute workloads for other satellites.

Next year, in October, our second satellite launching will have 10 times the capacity of the first one and will be the first commercial offering.

It's going to take about 3 years before we can compete with terrestrial data centres. Starcloud 2 is our intermediate satellite between here and there.

WT: Do you anticipate the competition with terrestrial centers to be a challenge?

Johnston: Terrestrial data centres use on the order of one million gallons of fresh water per day to cool the massive heat of the chips. One million gallons of fresh water reduces to zero water per day due to the average temperature of –235 degrees Celsius.

To keep pace with AI development, vast new data centers and many gigawatts of new energy projects to power them will need to be deployed around the world. At the same time, electrical utilities are being hit by a tidal wave of new demand from the electrification of industry, transport, and heating. Electricity demand may triple in the coming years as a result, but utilities in the Western world, hampered by planning restrictions, are not equipped for change at the required pace and scale. Without rapid adaptation, the upcoming energy crunch will hinder AI development.

Data centers in space can utilize high-intensity 24/7 solar power unhindered by day/night cycles, weather, and atmospheric losses (attenuation). This enables orders of magnitude lower marginal energy costs, resulting in drastic operating cost savings versus their terrestrial counterparts.

WT: Are launch costs a factor in the terrestrial/space competition?

Johnston: The world is on the verge of a step change in launch costs, thanks to the development of several partially or fully reusable heavy-lift launchers which are expected to offer a launch price of around $5 million per launch long term. With a payload capacity of 100 tons to Low Earth Orbit (LEO) Sun-Synchronous Orbit (SSO), this translates to approximately $30 per kilogram. It has been suggested that costs could drop to as low as $10 per kilogram. At these price points, launch costs are no longer a primary cost driver for orbital data centers.

WT: In conclusion, what would you like our viewers to focus on?

Johnston: Gigawatt-scale orbital data centers are among the most ambitious space projects of all time, sitting at the intersection of four trends: the drastic fall in launch costs, the upcoming electricity demand crunch, the growth in demand for large, energy-intensive GPU clusters, and the proliferation of low-cost connectivity from mega-constellations. We are convinced that orbital data centers are feasible, economically viable, and necessary to realize the potential of AI, the most important technology of the 21st century, in a rapid and sustainable manner It is very important to be conscious of the amount of water that data centres consume. Citizens in Canada and the US will benefit more from other uses of fresh water than processing data centres .

Related:

Starcloud video

White Paper