TAMPA, Fla. \u2014 Europe must move quickly to craft a large-scale strategy for space-based data centers or risk ceding a potential pillar of future digital infrastructure to global competitors, according to the European Space Policy Institute (ESPI).<\/p>\n
The independent think tank highlighted a surge of activity in the United States and China to advance in-orbit cloud computing, including growing interest from SpaceX\u2019s Elon Musk and other tech giants exploring space to ease artificial intelligence\u2019s soaring processing and energy demands.<\/p>\n
According to a report ESPI published Nov. 18, about 70 million euros ($81 million) of private capital has flowed into space-based data center ventures, or directly relevant enabling components, over the past five years.<\/p>\n
They include lunar data storage startup Lonestar of Florida and Washington-based Starcloud, which recently deployed its first small satellite<\/a>.\u00a0<\/p>\n Speculative discussions turn into tangible R&D<\/p>\n At around 60 kilograms, the Starcloud-1 spacecraft carries an Nvidia processor that is designed to run AI models in orbit, including variants of Google\u2019s Gemini.\u00a0<\/p>\n However, to scale into commercially meaningful capacity, Starcloud and many other orbital-processing ventures call for kilometers of solar arrays and giant radiators in orbit to shed gigawatts of heat.<\/p>\n This sheer scale highlights one of the most formidable engineering challenges in deploying full-scale orbital compute, ESPI noted. Despite major advances in recent years, launch affordability, thermal management and in-orbit assembly remain major barriers.<\/p>\n While Europe has produced early building blocks, from ESA\u2019s PhiSat AI processing missions to a feasibility study funded by the European Union<\/a>, ESPI pointed to how others are moving quickly toward operational systems.\u00a0<\/p>\n NASA manages more than two dozen AI-driven edge computing projects, including experiments on the International Space Station to process sensor readings in space, rather than sending raw data back to Earth.\u00a0<\/p>\n China, meanwhile, has launched the first 12 satellites of its planned Three-Body Computing Constellation<\/a>, envisioned as a 2,800-satellite AI computing mesh that performs heavy processing in space to reduce latency and downlink bottlenecks.<\/p>\n Powering AI<\/p>\n According to McKinsey analysts, as much as $6.7 trillion of investment<\/a> could be required in data centres by 2030, of which about $5.2 trillion is driven by AI workloads, if adoption and technology cycles continue on their current path.<\/p>\n \u201cThere is an increasing need in orbit for storage and processing in space for in-situ operations,\u201d Analysys Mason research director Claude Rousseau said via email, \u201cand to complement terrestrial-based data centre markets.\u00a0<\/p>\n \u201cGiven the strong demand from government customers for connectivity, it is a natural fit to launch space data centres in various forms to meet their growing needs for sovereignty of operations, coupled with an increase in use of AI and to strengthen data security.\u201d<\/p>\n Musk, who also leads the advanced-model startup xAI, said in October that an upcoming V3 generation of Starlink broadband satellites, set to fly on SpaceX\u2019s in-development Starship, could be scaled up to operate as orbiting data centers<\/a>, linked via high-speed lasers.<\/p>\n \u201cSpaceX will be doing this,\u201d he added.<\/p>\n