SpaceX, Blue Origin and startups race to launch AI data centers into orbit by 2027

Galactic Brain space datacenter promised in 2027

Getting inferencing infrastructure into orbit may soon be cheaper than building it down here Space startup Aetherflux says it plans to put its first data center satellite into orbit during the first quarter of 2027. The company, founded and run by Baiju Bhatt, co-founder of financial firm Robinhood, sees satellites as a time-saving alternative to terrestrial data center construction, which can take five or more years. "The race for artificial general intelligence is fundamentally a race for compute capacity, and by extension, energy," said Bhatt in a statement. "The elephant in the room is that our current energy plans simply won't get us there fast enough." Amid the baffling absence of funding constraints, efforts to scale artificial intelligence remain gated by the availability of data center capacity and energy. Mundane concerns like land acquisition, utility connections, and creating sturdy structures can be bypassed for the cost of lifting kit beyond the Kármán line. The cost to launch 1 kg on SpaceX's Falcon Heavy comes to about $1,400, according to recent estimates. Per Google's calculations, if launch costs drop to around $200 per kg, as projected by 2030, the outlay required to set up and run space-based data centers would be comparable to ground-based operations. Bhatt calls the satellite constellation project "Galactic Brain," though the solar-powered data satellites won't have such a vast remit - they'll just be orbiting the Earth alongside a growing number of other objects. Aetherflux joins Orbits Edge and Starcloud, not to mention Google and Nvidia, as companies with ambitions to put data centers in space, the final (minimally regulated) frontier now that data centers have been sunk into the ocean and buried underground. Bhatt's biz was founded in 2024 and scored $60 million of funding for the purpose of demonstrating the viability of beaming energy from space to Earth via infrared laser. It aims to launch a satellite capable of doing so in 2026. The company's space data center plans represent an expansion on its initial vision. Aetherflux isn't yet ready to talk about pricing. "Our first product will focus on AI inference general-purpose compute for customers," a company spokesperson told The Register by email. "We do not have a price point to disclose at this time." But for a to-be-determined fee, the company aims to provide "multi-gigabit level bandwidth with nearly constant uptime." When the company gets around to doing regular launches, it anticipates sending about 30 satellites up at a time on a SpaceX Falcon 9 or equivalent launcher. If SpaceX's Starship becomes an option, Aetherflux could orbit 100 or more datacenter sats in a single launch. Asked about the life expectancy of its birds, given that GPUs may not last more than a few years under high utilization and radiation, Aetherflux's spokesperson said: "Our approach is to continuously launch new hardware and quickly integrate the latest architectures. Older systems will run lower priority tasks and serve out the full useful lifetime of high-end GPUs." ®

'Greetings, earthlings': Nvidia-backed Starcloud trains first AI model in space as orbital data center race heats up

Starcloud's output Gemma in space. Gemma is a family of open models built from the same technology used to create Google's Gemini AI models. Starcloud wants to show outer space can be a hospitable environment for data centers, particularly as Earth-based facilities strain power grids, consume billions of gallons of water annually and produce hefty greenhouse gas emissions. The electricity consumption of data centers is projected to more than double by 2030, according to data from the International Energy Agency. Starcloud CEO Philip Johnston told CNBC that the company's orbital data centers will have 10 times lower energy costs than terrestrial data centers. "Anything you can do in a terrestrial data center, I'm expecting to be able to be done in space. And the reason we would do it is purely because of the constraints we're facing on energy terrestrially," Johnston said in an interview. Johnston, who co-founded the startup in 2024, said Starcloud-1's operation of Gemma is proof that space-based data centers can exist and operate a variety of AI models in the future, particularly those that require large compute clusters. "This very powerful, very parameter dense model is living on our satellite," Johnston said. "We can query, it and it will respond in the same way that when you query a chat from a database on Earth, it will give you a very sophisticated response. We can do that with our satellite." In a statement to CNBC, Google DeepMind product director Tris Warkentin said that "seeing Gemma run in the harsh environment of space is a testament to the flexibility and robustness of open models." In addition to Gemma, Starcloud was able to train NanoGPT, an LLM created by OpenAI founding member Andrej Karpath, on the H100 chip using the complete works of Shakespeare. This led the model to speak in Shakespearean English. Starcloud -- a member of the Nvidia Inception program and graduate from Y Combinator and the Google for Startups Cloud AI Accelerator -- plans to build a 5-gigawatt orbital data center with solar and cooling panels that measure roughly 4 kilometers in both width and height. A compute cluster of that gigawatt size would produce more power than the largest power plant in the U.S. and would be substantially smaller and cheaper than a terrestrial solar farm of the same capacity, according to Starcloud's white paper. These data centers in space would capture constant solar energy to power next-generation AI models, unhindered by the Earth's day and night cycles and weather changes. Starcloud's satellites should have a five-year lifespan given the expected lifetime of the Nvidia chips on its architecture, Johnston said. Orbital data centers would have real-world commercial and military use cases. Already, Starcloud's systems can enable real-time intelligence and, for example, spot the thermal signature of a wildfire the moment it ignites and immediately alert first responders, Johnston said. "We've linked in the telemetry of the satellite, so we linked in the vital signs that it's drawing from the sensors -- things like altitude, orientation, location, speed," Johnston said. "You can ask it, 'Where are you now?' and it will say 'I'm above Africa and in 20 minutes, I'll be above the Middle East.' And you could also say, 'What does it feel like to be a satellite? And it will say, 'It's kind of a bit weird' ... It'll give you an interesting answer that you could only have with a very high-powered model." Starcloud is working on customer workloads by running inference on satellite imagery from observation company Capella Space, which could help spot lifeboats from capsized vessels at sea and forest fires in a certain location. The company will include several Nvidia H100 chips and integrate Nvidia's Blackwell platform onto its next satellite launch in October 2026 to offer greater AI performance. The satellite launching next year will feature a module running a cloud platform from cloud infrastructure startup Crusoe, allowing customers to deploy and operate AI workloads from space. "Running advanced AI from space solves the critical bottlenecks facing data centers on Earth," Johnston told CNBC. "Orbital compute offers a way forward that respects both technological ambition and environmental responsibility. When Starcloud-1 looked down, it saw a world of blue and green. Our responsibility is to keep it that way," he added.

SpaceX Hypes 2026 IPO With Predictable Ingredient: AI in Orbit

SpaceX is reportedly preparing a 2026 initial public offering that would raise over $30 billion, targeting a valuation of about $1.5 trillion. Part of the funding is said to toward an ambitious venture that CEO Elon Musk and other tech billionaires have been eyeing for some time now: AI in space. As data centers proliferate across the U.S. to support surging demand for AI training and operation, the resulting strain on energy, water, and land resources is becoming impossible to ignore. Facing limits on the growth of this booming industry, Big Tech has proposed another solution: building data centers in orbit. Two sources familiar with SpaceX’s IPO plans told Bloomberg the company expects to use some of the new capital to develop orbital data centers, including purchasing the chips required to build them. Musk has expressed strong interest in the idea in recent weeks, claiming that SpaceX could leverage Starlink for this purpose. “Simply scaling up Starlink V3 satellites, which have high speed laser links would work. SpaceX will be doing this,†he posted on X in late October. During a recent event with Baron Capital, Musk said he sees a path to putting 100 gigawatts of computing power from solar-powered AI satellites into orbit each year. Gizmodo reached out to SpaceX for comment but did not receive a response by the time of publication. Proponents of orbital data centers say their advantage lies in tapping limitless solar energy while avoiding the environmental and public-health consequences of building them on Earth. Meanwhile, critics contend that the expense and technological complexity of putting data centers in the unforgiving environment of space makes the idea impractical. Many tech billionaires, including Musk, fall squarely into the “proponents†category. With a blockbuster IPO on the horizon and Musk's forecast that SpaceX's newest rocket will start launching Starlink V3 satellites in 2026 as well, the company does appear uniquely well-positioned to lay the groundwork for orbital server farms. It’s unclear exactly how SpaceX would adapt its Starlink architecture for AI data centers, or how much this venture would cost. It would be a major leap, but not an impossible one. Each Starlink V3 satellite is expected to have a downlink speed of 1 terabyte per second, a 10-fold increase from the V2 mini Starlink satellites, which is comparable to terrestrial data centers. It's also not yet clear how much processing power SpaceX can pack into each spacecraft. A large hurdle will be launching the many Starlink V3 satellites required to build the equivalent of a terrestrial data center. SpaceX can already launch thousands of satellites per year with its Falcon 9 rocket. But the V3s will be much bigger and will need to launch on Starship, which is not yet fully operational. Even if SpaceX hits the ground running next year, it could take decades to achieve this goal. SpaceX isn’t the only company interested in orbital data centers. Blue Origin founder Jeff Bezos has also touted them as a viable way to protect Earth’s resources and the AI industry’s growth, predicting that humans will start building gigawatt-scale orbital data centers within the next 10 to 20 years. In November, Google announced Project Suncatcher, a research initiative aimed at building its own satellite-based orbital data centers. Several smaller companies are also developing space-based data center technologies, including Aetherflux, Lonestar Data Holdings, and Axiom Space. The potential advantages of moving these facilities off the Earth have clearly captured the tech sector’s attention. Though SpaceX has not publicly confirmed plans for an IPO or for investment in orbital data center research and development, its reputation as an aerospace pioneer means these rumors are likely to spur competitors’ interest in taking AI infrastructure to space.

Billionaires Jeff Bezos, Elon Musk racing to build AI data centers in...

Jeff Bezos' aerospace firm Blue Origin has been working for over a year on the necessary technology for artificial intelligence data centers in space, the Wall Street Journal reported Wednesday, citing a person familiar with the matter. Meanwhile, Musk's SpaceX plans to use an upgraded version of its Starlink satellites to host AI computing payloads, pitching the technology as part of a share sale that could value the company at $800 billion, the report said, citing people involved in the discussions. SpaceX and Blue Origin did not immediately respond to requests for comment. Reuters could not immediately verify the report independently. The concept of orbital data centers has gained traction among tech giants as those on Earth have driven up demand for electricity and water to cool their servers. Amazon founder Bezos in October predicted that gigawatt-scale data centers would be built in space within the next 10 to 20 years and that continuously available solar energy meant they would eventually outperform those based on Earth. "We will be able to beat the cost of terrestrial data centers in space in the next couple of decades," Bezos said at the time. "These giant training clusters ... will be better built in space, because we have solar power there, 24/7. There are no clouds and no rain, no weather." Last week, Musk had dismissed media reports that SpaceX was raising funds at an $800 billion valuation, calling them inaccurate. SpaceX is looking to raise more than $25 billion through an initial public offering in 2026, a move that could boost the rocket-maker's valuation to over $1 trillion, Reuters reported on Tuesday.

Bezos' Blue Origin working on orbital data center technology, WSJ reports

Dec 10 (Reuters) - Jeff Bezos' aerospace firm Blue Origin has been working for over a year on the necessary technology for artificial intelligence data centers in space, the Wall Street Journal reported on Wednesday, citing a person familiar with the matter. Meanwhile, Musk's SpaceX plans to use an upgraded version of its Starlink satellites to host AI computing payloads, pitching the technology as part of a share sale that could value the company at $800 billion, the report said, citing people involved in the discussions. SpaceX and Blue Origin did not immediately respond to requests for comment. Reuters could not immediately verify the report independently. The concept of orbital data centers has gained traction among tech giants as those on Earth have driven up demand for electricity and water to cool their servers. Amazon founder Bezos in October predicted that gigawatt-scale data centers would be built in space within the next 10 to 20 years and that continuously available solar energy meant they would eventually outperform those based on Earth. "We will be able to beat the cost of terrestrial data centers in space in the next couple of decades," Bezos said at the time. "These giant training clusters ... will be better built in space, because we have solar power there, 24/7. There are no clouds and no rain, no weather." Last week, Musk had dismissed media reports that SpaceX was raising funds at an $800 billion valuation, calling them inaccurate. SpaceX is looking to raise more than $25 billion through an initial public offering in 2026, a move that could boost the rocket-maker's valuation to over $1 trillion, Reuters reported on Tuesday. (Reporting by Jaspreet Singh in Bengaluru; Editing by Alan Barona)