TZUTC: -0500   
   MSGID: 2097.consprcy@1:2320/105 2de89a8a   
   PID: Synchronet 3.21a-Linux master/123f2d28a Jul 12 2025 GCC 12.2.0   
   TID: SBBSecho 3.28-Linux master/123f2d28a Jul 12 2025 GCC 12.2.0   
   BBSID: CAPCITY2   
   CHRS: ASCII 1   
   FORMAT: flowed   
   Musk insists that 'the lowest cost way to generate AI compute will be in space'   
   within three years after SpaceX acquired xAI, but that timeline is more science   
   fiction than strategy   
      
   Opinion   
      
   The dream of orbital AI compute may come true someday, but don't bet on   
   Musk's clock   
      
   Elon Musk has added another line to his history of technological predictions   
   that sail far beyond optimistic and into the delusional. As part of announcing   
   the acquisition of his xAI company by (the also Musk-run) SpaceX, he declared   
   that not only was space ideal as a cheap location for running AI servers, but   
   that it would happen faster than most kitchen renovations on Earth.   
      
   "My estimate is that within 2 to 3 years, the lowest cost way to generate AI   
   compute will be in space," Musk wrote in the announcement. " This   
   cost-efficiency alone will enable innovative companies to forge ahead in   
   training their AI models and processing data at unprecedented speeds and   
   scales, accelerating breakthroughs in our understanding of physics and the   
   invention of technologies to benefit humanity."   
      
   Estimate is a term doing a lot of work here, because when you look closely, the   
   numbers don't add up, and neither does the physics. Still, it's a   
   headline-grabbing idea, now further amplified by SpaceX swallowing up xAI. The   
   idea of space-based AI processing isn't outlandish on its own. Other AI   
   developers have also been exploring the prospect, with both Google and Amazon   
   in initial design discussions. After all, AI is power-hungry, and space has   
   infinite sunshine and no water bills.   
      
   But a grand, interplanetary vision isn't the same thing as a realistic   
   business plan - especially not one that delivers within 36 months. The   
   infrastructure isn't ready. Merging an AI company with a rocket company   
   doesn't fast-forward the Earth's rotation. If you believe Musk will have AI   
   data centers in orbit before 2030, I've got a used Tesla humanoid robot to sell   
   you.   
      
   Imaginary booster rockets   
      
   Space offers uninterrupted solar radiation, ambient cold for thermal   
   dissipation, and the ultimate perk for remote work: zero zoning restrictions.   
   Musk's point isn't entirely unfounded. Data centers are energy-devouring   
   creatures, sucking up power, land, and water, and sparking political battles.   
      
   Meanwhile, in orbit, you're above the clouds and below the radar. No utility   
   bills. No water rights battles. There are many reasons to be intrigued by   
   orbital compute. But there are many more to be skeptical of its imminent   
   arrival.   
      
   Even assuming record-setting rocket launch schedules that are all successful,   
   getting mass to orbit still isn't cheap. Launching a full data center's   
   worth of equipment into space, with radiation shielding, thermal management,   
   fault tolerance, and redundancy, is not something that can be done affordably   
   in any timeline under a decade. And that assumes zero maintenance or upgrades.   
   Terrestrial centers swap out dead GPUs like old lightbulbs. Up there, your only   
   hope is robotic servicing or tons of redundancy.   
      
   And all that sunlight energy comes with plenty of less enticing radiation.   
   Cosmic rays, solar flares, and the general hostility of space are not side   
   issues. They're central to why most satellites are hardened, expensive, and   
   decades behind in chip design. GPUs built for inference and training are   
   fragile. They aren't designed to float above the Van Allen belt.   
      
   Not to mention the space trash. Putting thousands of compute satellites into   
   low-Earth orbit could cause a cascade of collisions. SpaceX is already dominant   
   in orbital traffic. Layering a second orbital network of AI computers could   
   raise significant regulatory and environmental backlash, even wittout constant   
   danger of crashes.   
      
   Decades, not years   
      
   As a long-term plan, space data centers could be a great option. They could   
   offload pressure from power grids, avoid zoning fights, and scale globally   
   without boiling local lakes. The physics aren't impossible, but the equations   
   translate to complex, difficult, expensive engineering. Three years for a   
   functioning AI data center in orbit is not serious, and people who say it will   
   happen shouldn't be taken seriously.   
      
   Not because people don't want to make the orbiting AI data centers happen,   
   but because large-scale infrastructure, especially in space, requires patience,   
   iteration, and a willingness to admit when Earth is still the better option.   
   Admitting mistakes and backing down from grandiose fever dreams are not habits   
   for Musk. But, like his robots, his fleet of self-driving cars, and his video   
   game prowess, the orbiting AI centers are laughable nonsense. Give the project   
   to real engineers and ask them about a real timeline, and we'll see how the   
   first satellites are doing in a decade or so.   
      
      
   https://www.techradar.com/ai-platforms-assistants/musk-insists-that-the-lowest-   
   cost-way-to-generate-ai-compute-will-be-in-space-within-three-years-after-space   
   x-acquired-xai-but-that-timeline-is-more-science-fiction-than-strategy   
      
   $$   
   --- SBBSecho 3.28-Linux   
    * Origin: Capitol City Online (1:2320/105)   
   SEEN-BY: 105/81 106/201 128/187 129/14 305 153/7715 154/110 218/700   
   SEEN-BY: 226/30 227/114 229/110 134 206 300 307 317 400 426 428 470   
   SEEN-BY: 229/664 700 705 266/512 291/111 320/219 322/757 342/200 396/45   
   SEEN-BY: 460/58 633/280 712/848 902/26 2320/0 105 304 3634/12 5075/35   
   PATH: 2320/105 229/426